Climates of the last three interglacials in subtropical eastern Australia inferred from wetland sediment geochemistry

Journal Pre-proof Climates of the last three interglacials in subtropical eastern Australia inferred from wetland sediment geochemistry

C.W. Kemp, J. Tibby, L.J. Arnold, C. Barr, P.S. Gadd, J.C. Marshall, G.B. McGregor, G.E. Jacobsen PII:

S0031-0182(19)30750-3

DOI:

https://doi.org/10.1016/j.palaeo.2019.109463

Reference:

PALAEO 109463

To appear in:

Palaeogeography, Palaeoclimatology, Palaeoecology

Received date:

12 August 2019

Revised date:

4 November 2019

Accepted date:

12 November 2019

Please cite this article as: C.W. Kemp, J. Tibby, L.J. Arnold, et al., Climates of the last three interglacials in subtropical eastern Australia inferred from wetland sediment geochemistry, Palaeogeography, Palaeoclimatology, Palaeoecology (2019), https://doi.org/10.1016/j.palaeo.2019.109463

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier.

Journal Pre-proof Climates of the last three interglacials in subtropical eastern Australia inferred from wetland sediment geochemistry

Kemp, C.W.1, Tibby, J.1, Arnold, L.J.2, Barr, C.1, Gadd, P.S.3, Marshall, J.C.4,5, McGregor, G.B.4, Jacobsen, G.E.3

North Terrace Campus, Adelaide, 5005, SA, Australia

of

1. Geography, Environment and Population and Sprigg Geobiology Centre, University of Adelaide,

ro

2. School of Physical Science, Environment Institute, Sprigg Geobiology Centre and Institute for

-p

Photonics and Advanced Sensing, University of Adelaide, North Terrace Campus, Adelaide, 5005,

re

SA, Australia

South Wales 2232, Australia

lP

3. Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, New

na

4. Queensland Department of Environment and Science, Dutton Park, Queensland, Australia

ur

5. Australian Rivers Institute, Griffith University, Nathan, 4111, Queensland, Australia

Christopher Kemp

Jo

Corresponding author details:

Email: [email protected]

Postal address: Department of Geography, Environment and Population, The University of Adelaide, North Terrace, Adelaide, South Australia, 5005

1

Journal Pre-proof Abstract

Records of Australian climate during Marine Isotope Stages 5 and 7 (130–71 and 243–191 ka) are rare, preventing detailed assessments of long-term climate, drivers and ecological responses across the continent over glacial-interglacial timescales. This study presents a geochemistry-based palaeoclimate record from Fern Gully Lagoon on North Stradbroke Island (also known as Minjerribah) in subtropical eastern Australia, which records climates in MIS 7a–c, MIS 5 and much of the Holocene, in

of

addition to MIS 4 (71–57 ka), and parts of MIS 6, MIS 3 and MIS 2 (191–130, 57–29 and 29–14 ka).

ro

Indicators of inorganic sedimentation from a 9.5 m sediment core – focussed on high-resolution

-p

estimates of sediment geochemistry supported by x-radiography, inorganic content and magnetic

re

susceptibility – were combined with a chronology consisting of six radiocarbon (14C) and thirteen single-

lP

grain optically stimulated luminescence (OSL) ages. Hiatuses occurred at ~178–153 ka, ~36–21 ka and ~7–2 ka and likely result from the wetland drying. Low values of locally sourced aeolian materials

na

indicate a wet MIS 7a–c and early MIS 6 before a relatively dry MIS 5. Inorganic flux during the Holocene was up to four times greater than during MIS 5, consistent with long-term interglacial drying observed in

ur

other regions, most notably in central Australia. This study highlights the importance of employing a

Jo

combination of multiple dating approaches and calibrated geochemical proxies to derive climate reconstructions and to identify depositional complexities in organic-rich wetland records.

Key words: Palaeoclimate, North Stradbroke Island, Holocene, MIS 5, MIS 7, µXRF

2

Journal Pre-proof 1. Introduction

Reliable reconstructions of past interglacial climates are important for providing analogues of future climate change, as well as for better understanding ecosystem responses that are likely to accompany increases in global temperature (Turney and Jones, 2010; Harrison and Bartlein, 2012). Records of interglacial climates are similarly required to understand the long-term evolution of Australian biota including past extinction events (Kershaw et al., 2003; Miller et al., 2016; van der Kaars

of

et al., 2017). At present, the usefulness of Australian palaeoclimate records for either of these purposes

ro

is relatively limited, as most only extend to the Last Glacial Maximum (LGM) ~20 thousand years ago (ka)

-p

(Petherick et al., 2013; Reeves et al., 2013).

re

Records of Australian climate during the past three interglacials are limited, but include Lynch’s

lP

Crater and the nearby ODP 820 marine record (Kershaw et al., 2007a; Moss and Kershaw, 2007) in the north-east; Lake Eyre, Lake Woods and central Australian streams and lakes (Bowler et al., 1998; Nanson

na

et al., 2008; Cohen et al., 2015; Fu et al., 2017); and the Naracoorte Caves, Lake Selina, Lake Wangoom

ur

and Caledonia Fen in the south-east (Ayliffe et al., 1998; Colhoun et al., 1999; Harle et al., 2002; Kershaw

Jo

et al., 2007b). The distribution of available records is limited in geographical scope, and does not currently include equatorial or subtropical Australian localities. Additionally, while these records infer past climate from proxies such as pollen, charcoal and palaeoshorelines, there is not as yet an Australian interglacial record utilising µXRF geochemical analysis.

Climate records derived from geochemical studies of wetland sediments have experienced a boom in the past decade, with the advent of relatively cheap, reliable, reproducable µXRF analysis (Croudace and Rothwell, 2015). Wetland geochemical records represent invaluable archives for palaeoclimatology reconstructions, providing constraints on climate change at local (e.g. Eggenberger et

3

Journal Pre-proof al., 2018; Vegas-Vilarrúbia et al., 2018; Burrows et al., 2016) and regional scales (e.g. Field et al., 2018; Pleskot et al., 2018; Profe et al., 2018), and enabling a greater understanding of internal wetland processes (e.g. Burrows et al., 2016; Kienel et al., 2017; Vegas-Vilarrúbia et al., 2018).

Here we present a multi-proxy ~209 kyr record of subtropical climatic variability from Fern Gully Lagoon focussing on past interglacials. This study is the foundation for future studies of regional climate change based on Fern Gully Lagoon sediments and assesses the validity of the site as a palaeoclimate

of

archive. We use calibrated µXRF Itrax scanning to infer locally-derived aeolian inorganic sedimentation

-p

ro

and assess the suitability of OSL and 14C to date the complex, organic-rich sediment sequence.

re

2. Background

lP

2.1 Global climate changes during recent interglacials

A recent review of global interglacial climates has demonstrated some important differences

na

between marine isotope stage 7 (MIS 7), MIS 5 and the Holocene (Berger et al., 2015). Specifically, the

ur

MIS 7 interglacial complex was characterised by higher average global temperatures than the MIS 5

Jo

interglacial complex but had a lower temperature peak in MIS 7e than that in MIS 5e (Lisiecki and Raymo, 2005; Parrenin et al., 2013; Berger et al., 2015). The lower MIS 7e temperature is likely due to the greater difference between MIS 5e summer and winter insolation, while insolation during MIS 7e was more evenly distributed over the seasons due to lower obliquity of the Earth’s orbit (Berger et al., 2015). In contrast to the warmer MIS 5e, the warmer climates of the MIS 7a–c sub-complex relative to the MIS 5a–c sub-complex are reflected in extended periods of reduced ice volume (Lisiecki and Raymo, 2005; Elderfield et al., 2012), higher sea levels (Rohling et al., 2009) and higher atmospheric CO2 concentrations (Köhler et al., 2017). MIS 7a–c represents a rare ~20 kyr period of relatively stable, warm, mean global climate (Berger et al., 2015). 4

Journal Pre-proof Sea surface temperature records from the West Pacific Warm Pool (WPWP) can be used to guide understanding of globally-linked, regional changes in eastern Australian climate that occurred during the previous three interglacials (Zhang et al., 2017). In particular, sea surface temperatures in the WPWP (Fig. 1) can dictate climate variability in the wider region, including the central-eastern Australian coast and the Tasman Sea (Martinez et al., 2002; Bostock et al., 2006; Pelejero et al., 2006). WPWP sea surface temperatures indicate that MIS 5e had the warmest SSTs of the last ~250 kyr (Tachikawa et al.,

-p

2.2 Australian climates during past interglacials

ro

interglacial complex by an average of ~1 – 2oC (Lo et al., 2017).

of

2014), but that the MIS 7 interglacial complex and the Holocene were/are warmer than the MIS 5

re

The available Australian MIS 7 data permits insights into broad-scale climatic trends. During late

lP

MIS 7, Lakes Eyre and Woods in central Australia, which are predominantly fed by monsoonal rains, were full, in contrast to their present mostly dry state (Bowler et al., 1998; Fu et al., 2017). Meanwhile,

na

wet climates and open water were present in Lynch’s Crater in ENSO-dominated north-eastern Australia

ur

(Kershaw et al., 2007a; Moss and Kershaw, 2007). Similarly, the speleothem record from the Naracoorte

Jo

Caves in south-eastern Australia, which primarily reflects southern westerly wind-dominated rainfall, indicates wet climates, with pronounced periods of calcitic growth between ~220 and 155 ka (Ayliffe et al., 1998). These few available records indicate that Australia was wetter during late MIS 7 than the Holocene, with a more active Australian Monsoon in the north and stronger or more northerly southern westerly winds.

Central Australia was also wet during MIS 5 (Bowler et al., 1998; Cohen et al., 2015), though the average precipitation:evaporation ratio was lower than during late MIS 7. North-eastern Australian records indicate wet climates during MIS 5, but greater climate variability than during MIS 7 (Kershaw et 5

Journal Pre-proof al., 2007a; Moss and Kershaw, 2007). The pattern in the south-east of the continent is in general agreement with north-eastern records, with generally wet climates observed during MIS 5 (Edney et al., 1990; Ayliffe et al., 1998; Colhoun et al., 1999; Harle et al., 2002).

Holocene hydroclimate in Australia was drier than MIS 5 and MIS 7 in most regions. Lake Eyre and Lake Frome, which were full during MIS 5, became dry and ephemeral (Cohen et al., 2015), while

of

speleothem growth largely ceased in the Naracoorte Caves (Ayliffe et al., 1998) suggesting weakening or

ro

more southerly south westerly winds. However, there is uncertainty about whether the Holocene was

-p

drier than MIS 5 in south-east Australia and the Wet Tropics of north-eastern Australia within each record (Edney et al., 1990; Colhoun et al., 1999; Kershaw et al., 2007a; Kershaw et al., 2007b).

re

Uncertainty within records combined with geographical separation makes it difficult to determine

lP

climate change at a continental scale. As such, additional records from new locations are required.

na

2.3 Study site

ur

North Stradbroke Island is a sand island located along the eastern coast of Australia (Fig. 1). The

Jo

Island is part of the world’s oldest and largest coastal dune system (Patton et al., 2019) and lies on an aeolian dust pathway from central and south-central Australia (McGowan et al., 2008; Petherick et al., 2009). The Island is a strategic location for understanding palaeoclimate, as it is situated in the subtropics (Fig. 1), with a contemporary climate that is strongly influenced by the El Niño – Southern Oscillation (ENSO) (Barr et al., 2019) and preserves the highest density of wetlands with sediment dating to the LGM in Australia (Tibby et al., 2017). As such, it has been the focus of several detailed palaeoenvironmental reconstructions (e.g. Moss et al., 2013; Barr et al., 2017; Petherick et al., 2017; Cadd et al., 2018), the longest of which extends to ~130 ka (Cadd et al., 2018). There is evidence for

6

Journal Pre-proof human occupation of the island from at least 20 ka at Wallen Wallen Creek (Fig. 1), with a Holocene increase in human occupation peaking at ~1 ka (Neal and Stock, 1986).

Fern Gully Lagoon (27.417°S, 153.460°E, 39 m ASL) is an approximately 0.8 km2 perched, palustrine wetland (Leach, 2011) that lies within a shallow bowl of vegetated dunes at the northwestern end of North Stradbroke Island. The wetland has two above-ground outflows, the largest being to the eponymous Fern Gully (Fig. 1c). The limited catchment area and highly permeable sandy soils

of

mean the delivery of fluvially transported material to the wetland is limited. Single-grain OSL dating of a

ro

reconnaissance core from Fern Gully Lagoon determined a preliminary basal age of ~208.4 ± 32.5 ka

-p

(Tibby et al., 2017) at a depth of 9.25 meters.

3.1 Core collection and correlation

lP

re

3. Methods

na

In 2015, two ~9 m-long cores were extracted from the approximate centre of Fern Gully Lagoon

ur

(27.4174°S, 153.4600°E) using a modified Bolivia corer, itself a modification of the Livingstone-square-

Jo

rod piston-sampling method (Wright, 1967). The two cores were offset by one metre horizontally and 50 cm vertically, in an attempt to provide a continuous sequence. Cores were extracted in black-painted PVC pipe and stored in black plastic sleeves, to eliminate light contamination and ensure the sediments would be suitable for luminescence dating. The first core (FG15-1) was chosen as the master core and was sampled for single-grain OSL and 14C dating. The second core (FG15-2) was used to fill gaps in the FG15-1 record and to provide additional sediment for 14C and conventional wave dispersive x-ray fluorescence analyses.

7

Journal Pre-proof Cores were aligned using CPLSlot sequence slotting software (Clark and Hounslow, 2009) using the eight elements with highest Itrax counts per second: silicon, titanium, zircon, potassium, calcium, iron and bromine. The correlation allowed the cores to be placed on a common depth scale.

Moisture and organic content were determined on 930 one cubic centimetre samples following Heiri et al. (2001). This record was used to target optimum OSL sample locations (i.e., locate areas with highest inorganic content, and presumably quartz content) and to calculate average water content and

-p

3.2 Itrax µXRF, magnetic susceptibility and X-radiography

ro

of

sediment densities for luminescence dose rate calculation (See supplementary materials).

re

To determine changes in sediment geochemistry, Itrax second-generation micro X-ray

lP

fluorescence (µXRF, Croudace et al., 2006) scanning of both cores was undertaken at the Australian Nuclear Science and Technology Organisation. Itrax scanning utilised a molybdenum x-ray tube at 30 kV

na

and 55 mA, with a 20 second exposure time, generating a record of elemental response (counts per second) for thirty-five elements at 2 mm intervals. An X-radiograph record using the same tube at 45 kV

Jo

ur

at 30 mA and a magnetic susceptibility record were also developed.

Micro XRF data must be normalised to mitigate the closed-sum effect (Weltje and Tjallingii, 2008; Löwemark et al., 2011). However, despite an increasing number of µXRF studies (Croudace and Rothwell, 2015), there is no standardised method for normalisation. To identify all recently published methods, we undertook a review and synthesis of µXRF in the Scopus (https://www.scopus.com/home.uri) and Web of Science (http://www.webofknowledge.com) databases. The results of this review were used to identify which normalisation methods have been utilised in recent µXRF studies, along with what (if any) validation and calibration methods were used (e.g. conventional XRF, inductively coupled plasma mass spectrometry (ICP-MS)). Further details of the 8

Journal Pre-proof database review methods and results, along with a discussion of recent µXRF methodologies are provided in the supplementary material.

The eight most common normalisation methods identified in our review were assessed to determine the best method for normalising the Fern Gully Lagoon µXRF record. These were: total counts per second normalisation (Bouchard et al., 2011; Martin et al., 2014), centred log-ratio normalisation (Weltje and Tjallingii, 2008; Weltje et al., 2015), aluminium normalisation (Brumsack, 2006; Löwemark et

of

al., 2011), titanium normalisation (Vegas-Vilarrúbia et al., 2018), silicon normalisation, zircon

ro

normalisation, normalisation using incoherent scattering (Compton scattering) and the ratio of

-p

incoherent to coherent scattering (Rayleigh and Compton scattering) (e.g. Guyard et al., 2007; Kylander

re

et al., 2011; Marshall et al., 2011; Berntsson et al., 2014).

lP

The normalised µXRF data were compared to quantitative conventional XRF analyses at twenty locations from core FG15-2, following a similar method to recent studies (e.g. Hahn et al., 2014; Falster

na

et al., 2018; Profe et al., 2018). The conventional XRF sample locations were chosen to represent the

ur

broadest range of µXRF-derived concentrations, but where there was little within-sample variability in

Jo

µXRF readings. The latter consideration was important since due to the high sediment organic content, it was necessary to sample up to 1 cm of the core to ensure sufficient material for XRF. Conventional XRF analysis was undertaken at the Commonwealth Scientific and Industrial Research Organisation in Adelaide, South Australia, using lithium-borate fusion of sample material. The glass discs created from the fusion procedure were subsequently analysed on a PANalytical Axios Advanced wavelength dispersive XRF (WD-XRF) system using an in-house silicates calibration program.

For normalisation methods without internal XRF calibration (all but centred log-ratio), µXRF counts per second were compared to absolute concentrations using simple linear regressions. The

9

Journal Pre-proof normalisation method with the highest coefficient of variation was used to identify the most suitable normalisation method for Fern Gully Lagoon sediments. For centred log-ratio, we used a Matlab script after Grant et al. (2017) to produce a multivariate centred log-ratio corrected dataset. This XRF calibration was then used to calculate the approximate mass of oxides for the composite record.

The commonly utilised µXRF incoherent/coherent scattering ratio was also evaluated as a qualitative indicator of sediment organic content (e.g. Guyard et al., 2007; Jouve et al., 2013; Field et al.,

of

2018; Woodward et al., 2018). µXRF water content corrections were also calculated after Boyle et al.

-p

ro

(2015), as the sediments have a high water content that is likely to affect µXRF readings.

re

3.3 Identification of the aeolian component in the Fern Gully Lagoon µXRF record

lP

McGowan et al. (2008) demonstrated that the inorganic component of wetland sediments on North Stradbroke Island (NSI) is derived from two main sources; the local silicon-rich and trace element-

na

depleted sands, and far-travelled dust, which is more clay-rich. Petherick et al. (2009) expanded on that study, analysing source material from 149 sites in central and eastern Australia to geochemically

ur

fingerprint far-travelled dust deposited in Native Companion Lagoon (Fig. 1b). High quantities of

Jo

scandium, gallium, thallium and nickel characterise the clay-rich, far-travelled sediment, contrasting with the Island’s rare earth element-depleted silicon-rich quartz sands (McGowan et al., 2008; Petherick et al., 2009). While scandium, gallium and nickel were identified in the µXRF scan, their concentration was too low to be accurate, and they were not considered further.

We derived a high-resolution record of calibrated µXRF silicon. Higher silicon input into wetlands is indicative of drier climate on NSI, which reduces protective vegetation cover (McGowan et al., 2008; Petherick et al., 2009). This effect is amplified by the sandy soils of NSI, which are generally loose and very dry due to their high permeability (Leach, 2011), which also restricts overland flow. 10

Journal Pre-proof To identify the main directions of variation in the Fern Gully inorganic data, a principal component analysis (PCA) was performed using the vegan package (Oksanen et al., 2018) in RStudio (Team-RStudio, 2015; Ihaka and Gentleman, 2016). A subset of seven normalised elements comprising those with the highest counts per second (silicon, iron, calcium, bromine, zircon, potassium and titanium), along with LOI estimated organic content were used. A broken stick model was used to estimate the number of significant components in the µXRF record (Fig. S2) (Baczkowski, 2000). The PCA

of

vectors were scaled by the species scores before being divided by the standard deviation of each

ro

element and multiplied by an equalising constant. In this way, element vectors were centred and

-p

standardised so that the relative variance of each element could be compared.

re

To characterise possible local sediment end members, samples were taken from the following



lP

locations:

The surface of modern dunes ~30 m from the edge of Fern Gully Lagoon and ~30 m from the

na

edge of nearby Welsby Lagoon.

Sand three meters below the surface of dunes surrounding Fern Gully Lagoon.



The surface of outflow one at Fern Gully Lagoon (Fig. 1c).



Local rock outcrops (Point Lookout rhyolite and Dunwich sandstone).



An indurated layer representing the B horizon of material below each of the two lagoons.

Jo

ur



These endmember samples were scanned using Itrax µXRF and placed on the PCA using the vegan predict model (Oksanen et al., 2018).

11

Journal Pre-proof 3.4 Dating

A preliminary age obtained from Fern Gully Lagoon exceeded the radiocarbon dating limit of ~50 ka (Tibby et al., 2017). As such, we employed two dating methods to examine different depths of the core in this study: 14C dating for sediments to a depth of 288 cm, and single-grain OSL dating from 171 cm to the base of the core. Four paired OSL and 14C samples were collected between 171 and 288 cm, to

of

assess age agreement.

14

ro

C dating was performed on 18 macrofossil samples, including seeds, leaf material, and bark

-p

and charcoal > 3 mm diameter. Six paired 14C samples were collected from different materials to identify any offset due to material properties, including absorption of humic acids and alteration by ground

re

water residence time (Hofmann et al., 2019). Sample preparation was undertaken at the Australian

lP

Nuclear Science and Technology Organisation, where each of the samples was dissected under a stereoscope. Any identified root fragments or other foreign objects were removed. Each sample then

na

underwent an Acid-Base-Acid (ABA) pre-treatment to remove humic acids, following Brock et al. (2010)

ur

and were freeze-dried before graphitisation and AMS measurement. After correction for isotopic

Jo

fractionation using measured δ13C values, the conventional 14C ages were calibrated using the Southern Hemisphere Calibration Curve (SHCal13; Hogg et al. 2013).

Single-grain OSL dating was performed on 19 quartz samples collected from core sections with the highest inorganic content. The sediment samples were extracted under filtered and subdued red LED lighting, where ~12 g of bulk sediment (dry weight) was retained from the exposed core face and margins for beta and gamma dose rate determination and beta dose rate water correction (after oven drying at 100oC). Additional bulk sediment was collected from the overlying and underlying 10 cm depth of each OSL sample position for gamma dose rate determination and gamma dose rate water content

12

Journal Pre-proof correction. Single-grain equivalent dose (De) measurements and environmental dose rate assessments were made using the experimental procedures described by Demuro et al. (2013; 2015). Further details of the OSL methodologies employed in this study, including SAR suitability assessments (dose recoveries) and quality assurance criteria, are included in the supplementary materials.

The age-depth model for Fern Gully Lagoon was constructed from the 14C and OSL likelihoods using a Bayesian Poisson process depositional model in OxCal v4.2.4 (P_Sequence model: Bronk-

of

Ramsey, (2009); Bronk-Ramsey and Lee, (2013)), which allows for randomly variable deposition rates

ro

through the age-depth profile. The P_Sequence k0 base rigidity parameter, which controls the ability of

-p

the model to respond to variations in the prior and likelihood data, was set to a single event per 1 cm of

re

sedimentation but was allowed to vary between 0.01 to 100 events per centimetre to accommodate any

lP

major fluctuations in deposition rate. This additional flexibility is at the expense of precision in the final age model and results in a liberal estimate of uncertainties inherent in the data.

na

As part of the assessment of modelling priors, we used a common method for identifying

ur

change in linear records: pruned exact linear time (PELT) multi-changepoint analysis (Killick and Eckley,

Jo

2014). PELT was performed on the sediment inorganic content record and the µXRF Si, Ti, K and Zr data to identify any statistically significant changes in deposition mode, including potential sedimentation hiatuses (see supplementary material for further details and results). These independently assigned changepoint locations were used to identify separate depositional units in the OxCal modelling framework (Fig. 3, S3), each of which was represented by a separate P_Sequence with delineating start and end boundaries, nested within a master Sequence according to stratigraphic priors. Posterior dated events were automatically calculated at 1 cm intervals throughout the sequence. Further details of the Bayesian age-depth modelling method are included in the supplementary materials.

13

Journal Pre-proof 4. Results

4.1 Core correlation

Sequence slotting revealed that the top of the two cores were vertically offset by ~97 cm, rather than the 50 cm planned (Fig. 2). The unplanned offset meant that a total of 15 cm of sediment located at the end of several core drives was not recovered, with the largest contiguous gap being 4 cm. These lost

of

sections did not coincide with the boundaries defined using change-point analysis.

-p

ro

4.2 µXRF normalisation and calibration

Eight of the twelve elements from the µXRF Fern Gully Lagoon sequence had significant

re

correlations (at p < 0.05) with the WD-XRF samples when normalised by total counts per second, with

lP

the four best correlated elements returning a coefficient of determination of r2 > 0.8, p < 0.05 (Table 1, Fig. S1). Normalisation via titanium, silicon or zircon was largely unsuccessful (r2 < 0.5). Of the remaining

na

21 recorded elements not included in WD-XRF regressions, 14 could not be considered representative of

ur

elemental concentration either due to low mean counts per second (Striberger et al., 2010; van der Bilt

Jo

et al., 2015) or having >5% of data with zero values (Sáez et al., 2009).

A lack of correlation between the incoherent/coherent scattering ratio and measured water content (r2 = 0.05, p < 0.05, n = 804) indicated that it was not possible to correct the µXRF record for water content (see Boyle et al., 2015). However, calibration of the µXRF record by conventional XRF accounts for any offset due to water content. Inorganic content estimated by the incoherent/coherent scattering ratio, as used in many recent studies (e.g. Guyard et al., 2007; Jouve et al., 2013; Burrows et al., 2016; Mackenzie et al., 2017), had a very weak relationship (r2 = 0.09, p < 0.05) to inorganic content estimated by LOI, possibly due to the very high organic content of the core sequence.

14

Journal Pre-proof 4.3 Fern Gully Lagoon sediment composition

Fern Gully Lagoon sediments are black, finely grained and highly organic, with some macrofossils, such as roots and wood fragments. They have an average of only ~8% inorganic material by dry mass (Fig. 3). Water content is also high, at 36–93% of wet sample weight. Sediment density and water content remained relatively constant down core, with no overall monotonic trend (Fig. 3),

of

demonstrating that there is little compaction of the sediments with increasing overburden pressure.

ro

Both inorganic content and µXRF-derived silicon content exhibit several peaks and a plateau of

-p

~10% inorganic content from ~170–270 cm (Fig. 2). Lower zircon levels in this part of the record likely indicate smaller average grain size (Cuven et al. 2010). There was a strong correlation (r2 = 0.70, p <

re

0.05) between the LOI inorganic content and µXRF-derived silicon, indicating quartz sand is the major

lP

inorganic component of the sediment. While the magnetic susceptibility and X-radiograph records indicate variability in the lower part of the record, aligning with photographed banding (Fig. 3), the

na

inorganic content and the µXRF silicon records indicate very little change, with largely consistent low

Jo

4.4 14C dating

ur

quantities of inorganic material present.

The AMS 14C ages have a large amount of scatter (Table 2, Fig. 4). In general, the 14C ages obtained from identifiable terrestrial macrofossils (gum nuts, seeds, leaves and bark fragments) are stratigraphically consistent over the uppermost 2 m of the core (n = 7) (Table 2). Three samples collected from below one metre yield very young outliers: sample RC-3 from 115 cm has a modern 14C age, and samples RC-7 and RC-14 from 159 cm and 215 cm, respectively, produced ages of <2 cal. ka. These three samples were originally identified as reed stalk segments but were subsequently inferred to be root fragments. This re-interpretation was based on roots being identified during further sampling in 15

Journal Pre-proof the same core sections. Additional evidence for localised root penetration is apparent from some of the single-grain OSL De distributions (section 4.5). Encouragingly, there is a broad agreement between the paired radiocarbon ages obtained at the same depths using terrestrial seeds and terrestrial leaf fragments (Table 2). However, there is an age offset of ~6 kyr for the replicate bark and reed stalk samples collected from a depth of 79 cm. This offset could again be attributed to misidentification of root material at this depth.

of

Five 14C samples were single >3 mm charcoal fragments. 14C dating of charcoal can underestimate

ro

sample age in organic-rich wetlands unless prepared using a specific method to remove humic acid

-p

contamination (Nilsson et al., 2001; Turetsky et al., 2004; Brock et al., 2011). Contamination is

re

particularly an issue for old samples where small concentrations of humic acid can produce large

lP

underestimations (e.g. 1% modern carbon contamination can cause a 15 kyr underestimation in a 50 ka sample) (Brock et al., 2011). As the charcoal samples from Fern Gully Lagoon were prepared using ABA,

na

the resultant ages may have been inadvertently affected by humic acid contamination (Brock et al., 2010). Indeed, there is some evidence to suggest this might be the case for the three charcoal samples

ur

collected from depths of 228, 235 and 288 cm, which have statistically indistinguishable 2σ calibrated

4.5 OSL dating

Jo

age ranges (Fig. 4, Table 2).

The single-grain OSL dating results are summarised in Table 3, with representative single-grain De distributions shown as radial plots in Figure 5. The OSL samples from Fern Gully Lagoon exhibit a broad range of De distribution characteristics, indicative of spatially and temporally variable bleaching and mixing conditions at the site (e.g., Arnold and Roberts, 2009; Arnold et al., 2007, 2008, 2012). Full

16

Journal Pre-proof discussions of the single-grain De distributions and statistical age models used to derive representative burial dose estimates for each sample are provided in the supplementary materials.

The final OSL ages exhibit good stratigraphic consistency for thirteen of the nineteen samples (Table 3). The remaining six samples (FG15 OSL-1, -8, -8-2, -9, -10 and -16) yield very young, outlying and inverted ages of between 6.69 ± 2.0 ka to 21.2 ± 2.6 ka for the lowermost 7 m of the core sequence. These outlying ages in the lowermost 7 m of the core are in keeping with the complex De distribution

of

characteristics observed for the six samples, which are characterised by very high overdispersion values

ro

(53–119%) and multiple discrete dose components (Fig. S7). These multimodal De datasets are

-p

interpreted as reflecting the presence of locally intruded young grain populations. These populations

re

could potentially be the result of sporadic lake desiccation and the formation of deep surface cracks (a

lP

process that is visible today in analogous peat-rich wetlands on NSI) and, or, downward grain transportation via root penetration into older sediments, as has been found previously (e.g. Bateman et

na

al., 2007; Brill et al., 2012). These interpretations are consistent with the presence of modern or near-

ur

modern organic remains at depths >1 m in the 14C study (section 4.4).

Jo

Assuming that the multiple discrete dose components of FG15 OSL-1, -8, -8 2, -9, -10 and -16 can be explained by localised post-depositional mixing, it follows that the bulk (sample-average) dose rate of these six samples may not be entirely representative of that experienced by either dose component during burial. Owing to the impracticalities of retrospectively deriving a component-specific dose rate for the multiple identified components, these samples are not considered suitable for dating. The ages shown for these samples in Table 3 are included for indicative purposes only and have not been included as likelihood estimates in the Bayesian age model.

17

Journal Pre-proof 4.6 Bayesian age-depth model

Thirteen single-grain OSL ages and eighteen 14C ages were included in the Bayesian age modelling procedure, separated by three depositional boundaries. A preliminary version of the model tested using these 31 likelihood estimates failed to converge owing to the identification of major statistical outliers. We excluded all potentially inaccurate charcoal and misidentified reed stalk ages from the second model (see Table 2). The youngest charcoal 14C sample (FG15 RC-6) had an age that was

of

consistent with the OSL sample collected at 175 cm depth (FG15 OSL-18) (see Section 4.2), suggesting

ro

that it may not be affected by humic acid contamination. However, for the sake of consistency, it was

-p

excluded from further consideration. These quality control measures resulted in only seven of the

re

original eighteen 14C likelihood estimates being included in the second Bayesian age model. A preliminary version of this model failed to converge owing to the identification of a major statistical

lP

outlier (sample FG15 RC-15). It was, therefore, necessary to eliminate this 14C likelihood estimate in the

na

final Bayesian model.

ur

The Bayesian model for Fern Gully Lagoon has Amodel and Aoverall values of 62.3% and 64.6%,

Jo

respectively, marginally exceeding the minimum acceptance threshold of 60% and thus indicating a valid model (Bronk Ramsey, 2009). The modelling results are summarised in Table S4 and illustrated in Fig. 6. The final model indicates a basal age of 209.3 ± 28.4 ka (±2σ) at 900 cm, in agreement with the previously measured basal age of 208.4 ± 32.5 ka on an adjacent preliminary core (Ad13069, Tibby et al. (2017)). The model reveals four distinct sedimentation phases separated by hiatuses: a late MIS 7 to early MIS 6 phase (209.3 ± 28.4 ka to 177.5 ± 25.4 ka, 930 – 288 cm), a late MIS 6 to late MIS 3 phase (155.2 ± 24.9 to 34.7 ± 14.5, 288 – 166 cm), a mid-MIS 2 to mid-Holocene phase (20.9 ± 1.4 ka to 6.5 ± 5.6 ka, 166 – 52 cm) and a late Holocene phase (1.7 ± 1.0 ka to 0.45 ± 0.4 ka, 52 – 0 cm) (±2σ, Table S4,

18

Journal Pre-proof Fig. 6). The average sedimentation rates of these four phases are (from bottom to top) 0.20 m/kyr, 0.01 m/kyr, 0.08 m/kyr and 0.31 m/kyr.

4.7 Identifying aeolian and autochthonous wetland sediment

Broken stick analysis revealed two significant principal components in the µXRF data (Fig. S2), which combined reflect greater than 97% of the variance in the data. Co-varying silicon, titanium,

of

potassium and zircon defined much of PC1. PC1 is very similar to the local aeolian signature identified by

ro

McGowan et al. (2008), and we, therefore, associate PC1 with local aeolian input. PC2 was defined by

-p

iron and calcium and is likely associated with a secondary aeolian source or internal wetland processes (Fig. 7). Unfortunately, the position of iron and calcium compounds within the sediment may have been

re

altered post-sedimentation, such as by changing redox, rendering the identification of a source material

lP

difficult. Inorganic MIS 5 and Holocene sediments from Fern Gully Lagoon are most similar to contemporary local dune surface and subsurface samples, while the closest source material to MIS 7

na

inorganics is Dunwich sandstone, which contains notable quantities of iron and calcium.

Jo

ur

4.8 Inorganic sedimentation at Fern Gully Lagoon

The influx of aeolian inorganics was highly variable over the past 209 ka (Fig. 8). Aeolian inorganic input during MIS 7a–c and early MIS 6 (209.3 ± 28.4 ka to 177.5 ± 25.4 ka) was low while other inorganic sedimentation (represented by calcium and iron dominated inputs) gradually declined. The earliest phase had a high rate of total sedimentation, but no notable inorganic peaks. During the late MIS 6 to late MIS 3 phase (155.2 ± 24.9 to 34.7 ± 14.5 ka), there were aeolian inorganic peaks at ~144 ± 10 ka (LOI, magnetic susceptibility, PC1), ~100 ± 8 ka (LOI, PC1) and ~52 ± 11 ka (PC1). During the midMIS 2 to mid-Holocene phase (20.9 ±0.8 to 6.5 ± 2.8 ka), there were some small inorganic peaks from ~14–11 ka. Aeolian inorganic sedimentation peaks during the late Holocene, with ~95% inorganic 19

Journal Pre-proof sediments recorded at ~1.6 ± 0.5 ka (magnetic susceptibility, LOI, PC1). While wetland desiccation and cracking during dry periods, indicated by OSL De distributions, may have displaced quartz sand grains from their original position in the sediment, it is unlikely that this had more than a minor influence on the aeolian inorganic record, as most sediment would have remained intact.

5. Discussion

of

5.1 µXRF sediment analysis of highly organic wetlands

ro

The most commonly used µXRF normalisation method (Table S4) – normalisation by total counts

-p

per second – was the most suitable for calibrating inorganic elements in the highly organic Fern Gully

re

Lagoon sediments. However, raw counts per second had comparable or better performance than most normalisation methods, indicating that raw data may be used in some circumstances (Table 1). The best

lP

performing method for correcting µXRF data appeared to depend on the element(s) under

na

consideration. For example, centred log-ratio correction resulted in the best correlation to WD-XRF particularly for magnesium, bromine and calcium (r2 = 0.85, r2 = 0.84 and r2 = 0.78 respectively, where p

ur

< 0.05, n = 20 for all values), while multiple normalisation methods for aluminium had similarly good r2

Jo

values (0.62 – 0.69, Table 1). For µXRF analysis of highly organic sediments, the optimal method may be to determine the most suitable normalisation method for each element before calibration, rather than using a single normalisation method. If testing of multiple normalisation methods is not possible, centred log-ratio normalisation will produce the best results in the majority of cases, as it has been mathematically proven to counter the closed sum and other distorting effects inherent in µXRF data (Weltje and Tjallingii, 2008; Croudace and Rothwell, 2015).

Several recent palaeoclimate reconstructions from wetland sediments have relied on µXRF derived iron (e.g. Rees et al., 2015; Burrows et al., 2016; Stephens et al., 2018) or magnesium (Foerster

20

Journal Pre-proof et al., 2018) records to infer, for example, waterlogged soils, redox conditions and detrital input. However, our analysis indicates that these elements may not be accurately characterised by µXRF in highly organic sediment. Bromine was used as an indicator of organic content in the Paddy’s Lake record (north-western Tasmania), supported by the incoherent to coherent ratio (inc/coh) (Beck et al., 2017). However, while inc/coh and bromine correlate, neither correlated well with organic content in our study (r2 < 0.003, p = 0.132, n = 804 and r2 = 0.05, p < 0.05, n = 804 respectively). Indeed, the use of inc/coh

of

requires additional stages of calibration to accurately indicate organic content (Woodward and Gadd,

ro

2019). While some recent studies have validated the use of the µXRF inc/coh for estimating organic

-p

content (e.g. Field et al., 2018; Woodward et al., 2018), a number have not (e.g. Rees et al., 2015; Turner et al., 2015; Burrows et al., 2016; Pleskot et al., 2018). Standardising normalisation and calibration of

re

µXRF records as a minimum requirement for µXRF derived climate studies would improve record

lP

precision and reduce uncertainty in future work.

na

5.2 Dating wetland sediments

ur

Reliable dating of Australian palaeoclimate sequences that span multiple past interglacial periods has proved difficult, although luminescence dating offers a potentially useful means of filling chronological

Jo

gaps across a range of palaeoenvironmental contexts (Fu et al., 2017; Roberts et al., 2018; De Deckker et al., 2019). This study highlights the complexities that can be encountered when using single-grain OSL and 14C dating in organic-rich peaty wetlands. By employing both techniques, it has been possible to better diagnose complications related to material selection (e.g., root fragments and charcoal), postdepositional mixing, and heterogeneous bleaching of sand grains. While a relatively high proportion of the original dating samples were unsuitable for inclusion in the final age model (i.e. 67% of the eighteen 14

C samples and 32% of the nineteen OSL samples), our results highlight the general suitability of 14C and

OSL dating in this depositional context when targeting optimal sampling materials, laboratory protocols 21

Journal Pre-proof and scales of analysis. Careful consideration of dating quality control has proved critical for deriving a meaningful age-depth model at Fern Gully Lagoon. There is good scope for using systematic single-grain OSL dating studies to refine the chronology of other Australian interglacial records.

5.3 Fern Gully Lagoon sediment hiatuses

Age-depth modelling indicates that there were three hiatuses in sedimentation at ~177 to 155

of

ka, ~34.7 to 20.9 ka and 6.5 to 1.7 ka. Palaeoclimate reconstructions from nearby wetland sites on North

ro

Stradbroke and Fraser Islands have revealed sedimentary hiatuses during dry periods such as the LGM

-p

(Donders et al., 2006; Woltering et al., 2014). Unfortunately, there are no nearby records of sufficient age and resolution to allow comparison of pre-MIS 5 hiatuses at FGL. The timing of two hiatuses in the

re

Fern Gully Lagoon record during mid-MIS 6 and late MIS 3 to mid-MIS 2 are consistent with drier

lP

climates observed in a number of Australian records, both in north-east (Kershaw et al., 2007a; Moss and Kershaw, 2007) and central Australia (Fu et al., 2017). The start of the mid-MIS 3 sedimentary hiatus

na

in Fern Gully Lagoon aligns with the commencement of widespread drying of Australia (Kemp et al.,

ur

2019), and occurs at the same time as drying in nearby Welsby Lagoon (Cadd et al., 2018). The mid-

Jo

Holocene hiatus at Fern Gully Lagoon has no equivalent from Welsby Lagoon (Cadd et al., 2018), although Fern Gully Lagoon may be more hydrologically sensitive due to its smaller catchment. Lower rainfall was noted at Swallow Lagoon after ~3 ka (Barr et al., 2019), while mid-Holocene hiatuses or drier phases were recorded at Hidden Lake and Lake Allom on nearby Fraser Island (Longmore, 1997; Donders et al., 2006).

The absence of a hiatus at Welsby Lagoon during the late Holocene after ~3 ka may possibly be due to a fire which burned Fern Gully Lagoon but not Welsby Lagoon. Loss of wetland peat due to fire may exceed 50 cm in a single event (in which a peatland may burn for more than a month) – high peat

22

Journal Pre-proof density and peat moisture content lower than 16% contribute to the loss of greatest material (Ballhorn et al., 2009; Davies et al., 2013; Lukenbach et al., 2015). However, an initial study of sedimentary charcoal did not indicate higher levels during this period (Kemp, unpublished data), and there is no notable change in the calcium record (Fig. 2) which could indicate mineralisation within the sediment as the result of a major peat fire (Smith et al., 2001). Therefore, it is more likely that Fern Gully Lagoon is more hydrologically sensitive than Welsby Lagoon, and drying led to a late-Holocene hiatus, rather than

of

loss of peat due to a major fire event.

-p

ro

5.4 Palaeoclimate interpretation and comparison with other records

We interpret increased aeolian inorganic sedimentation in Fern Gully Lagoon as indicative of dry

re

climates, as reduced vegetation cover increases local wind erosion (McGowan et al., 2008). While

lP

changes in wind strength could also be a major source of changing inorganic flux, records of terrestrial dust grain size from the Tasman Sea indicate that regional wind strength was secondary to continental

na

drying in explaining regional aeolian inorganic transport (Hesse and McTainsh, 2003). Increased biomass

ur

burning on NSI is a possible additional driver for increased inorganic sedimentation. However,

Jo

macrocharcoal and the proportion of sediment inorganic matter are uncorrelated at Fern Gully Lagoon (r2 = 0.04, p > 0.05, n = 632 (Kemp, unpublished data)), a situation similar to nearby Welsby Lagoon (Barr et al., 2017). Hence, it appears that increased biomass burning did not play a large role in increasing wind erosion on NSI. Sea-level transgression does not appear to drive inorganic flux to Fern Gully Lagoon as the major marine transgressions during late MIS 6 and after the LGM do not coincide with increased inorganic flux (Fig. 8). However, the lack of inorganic sediment immediately after the LGM may also be due to a change in the dominant regional wind direction which deposited sands offshore as postulated by Walker et al., (2018).

23

Journal Pre-proof We compared our results to the ~33 kyr inorganic flux record from Native Companion Lagoon (McGowan et al., 2008; Petherick et al., 2008) (Fig. 8) and the ~37 kyr record from Tortoise Lagoon (Petherick et al., 2017). The NCL record indicates that a dry phase occurred during the LGM as well as an increase in inorganic sedimentation at ~3 ka (McGowan et al., 2008). While Tortoise Lagoon records a similar LGM peak, there is no increase late-Holocene inorganic sedimentation (Petherick et al., 2017). The LGM inorganic peak was not recorded at Fern Gully Lagoon due to a hiatus, the LGM hiatus itself

of

indicates that Fern Gully Lagoon was similarly dry.

ro

The Fern Gully Lagoon sequence recorded low aeolian sedimentation with low PC1 scores and

-p

the lowest inorganic content (~3%), during MIS 7a–c, indicating that it was likely the wettest interglacial

re

of the past three in subtropical eastern Australia. This finding is similar to the record of moisture

lP

availability from central Australia, where recorded lake levels reached their highest for the past ~250 ka (Fu et al., 2017); Lynch’s Crater in north-east Australia, where there were wet climates with open water

na

present (Kershaw et al., 2007b), and Naracoorte caves in south-east Australia, where pronounced periods of calcitic growth occurred (Ayliffe et al., 1998). In comparison, there was greater aeolian input

ur

during MIS 5, indicating a drier local climate than in MIS 7a–c (Fig. 8; MIS 5 average ~0.24 kg/m2/kyr vs

Jo

average ~0.14 kg/m2/kyr inorganics). The record also contains a notable MIS 5 dry period at 100 ± 15.6 ka. A similar comparatively drier mid-MIS 5b–d from ~110–90 ka has been observed in other regions, most notably in tropical north-east Australia (Kershaw et al., 2007a). Conversely, central Australian lakes maintained shallow to deep-water conditions during MIS 5 (Bowler et al., 1998, Fu et al., 2017), while calcite growth occurred after ~105 ka (Ayliffe et al., 1998).

During the Holocene, inorganic aeolian deposition at Fern Gully Lagoon averaged ~0.96 kg/m2/kyr, four times greater than during MIS 5 and almost seven times MIS 7 levels. Greater inorganic flux, in combination with high PC1 scores, suggests the Holocene is the driest interglacial of the last 24

Journal Pre-proof three for subtropical eastern Australia. A drier Holocene relative to MIS 7a–c and MIS 5 is observed in several other Australian records, either via increased dry forest and herbaceous vegetation (Harle et al., 2002; Moss and Kershaw, 2007) or lower flows and lake filling (Maroulis et al., 2007; Nanson et al., 2008; Fu et al., 2017). Holocene drying has been attributed to increasing El Niño frequency compared to past interglacials (Moss and Kershaw, 2007), or to the extended period of high sea levels during the Holocene limiting WPWP influence on monsoonal precipitation and the warm East Australian Current (Nanson et

of

al., 2008). Holocene inorganic sedimentation at Fern Gully Lagoon may also be influenced by human-

ro

induced biomass burning. There is a long record of human activity on the island (Neal and Stock, 1986)

-p

which may have also contributed to mobilisation of dune sands. However, the paucity of dated archaeology close to Fern Gully Lagoon currently precludes an assessment of human influence on the

lP

re

Fern Gully record.

The pattern of increasingly dry interglacials observed at Fern Gully Lagoon since MIS 7 is

na

consistent with the few other Australian records covering this period, most notably in central Australia (Bowler et al., 2001; Cohen et al., 2015; Fu et al., 2017), but also in the north (Bowler et al., 1998) and

ur

south-east (Edney et al., 1990) of the continent. However, records from eastern Australia – tropical

Jo

Lynch’s Crater (Kershaw et al., 2007a) and alpine, temperate, Caledonia Fen (Kershaw et al., 2007b) – indicate similar climates during the Holocene and MIS 5. Similar climates between the north-eastern Lynch’s Crater and Fern Gully Lagoon may be explained by the behaviour of the East Australian Current (EAC). SST records from the Tasman Sea, which record the passage of the EAC, indicate similar temperatures for MIS 5e and the Holocene (Kawagata, 2001) as a result of more southerly penetration of the EAC. The more southerly penetration of the EAC may, in turn, may have driven convective rainfall in subtropical and temperate eastern Australia somewhat independent of major continental climate drivers such as ENSO.

25

Journal Pre-proof 6. Conclusions

Analysis of a new discontinuous sedimentary record from Fern Gully Lagoon using 14C dating, single-grain OSL dating and µXRF core scanning has enabled the reconstruction of a regional palaeoclimate sequence spanning the last three interglacial complexes. An evaluation of µXRF normalisation methods indicated that normalisation by total counts per second was the best method for

of

Fern Gully Lagoon sediments.

ro

The elemental signature of increased aeolian input to Fern Gully Lagoon, associated with drier

-p

climates, was characterised by PCA. The record indicates a relatively wet MIS 7a–c and early MIS 6 phase, a relatively drier MIS 5 interglacial complex, MIS 4 and MIS 3, and a wetter late MIS 2 which

re

transitioned into a drier Holocene. There is general agreement between the Fern Gully Lagoon record

lP

and central Australian lake records fed by the Australasian monsoon and records from the north- and south-east of the continent. Differences between ENSO driven north-eastern Australian records, and

na

Fern Gully Lagoon may be due to the influence of the East Australian Current.

ur

The increasingly dry interglacials observed from Fern Gully Lagoon may be due to more frequent

Jo

El Niño events. Understanding how ENSO changes during interglacials is important in predicting future water availability due to climate change, as well as ecosystem response to increasing global temperatures. However, due to limited climate records, further study is required. The Fern Gully Lagoon record, which is sensitive to changing ENSO, may assist in answering some of these questions.

It is difficult to isolate a detailed record of hydrological change without considering changing vegetation types and biomass burning records. Ongoing analysis of multiple climate proxies such as pollen, charcoal and stable isotopes from Fern Gully Lagoon will likely result in a greater understanding of past interglacial climates and their drivers. 26

Journal Pre-proof 7. Acknowledgements

We acknowledge Minjerribah (NSI) and the surrounding waters as Quandamooka Country, and we thank the Quandamooka Yoolooburrabee Aboriginal Corporation for their support of this research. This project was supported by the Australian Research Council Discovery Project DP150103875 and Future Fellowship Project FT130100195. Radiocarbon dating and Itrax µXRF scanning were achieved with the support of AINSE grant ALNGRA16003. We would like to thank Cameron Schulz (DES) and Dr

of

Harald Hofmann (UQ) for field assistance, and the Queensland Government for the Fern Gully Lagoon

ro

topographic map. We would also like to thank Mark Raven for WD-XRF analysis and Dr Katherine Grant

-p

for support running centred log-ratio calibration on Itrax data, as well as Chester Willard for his useful

Jo

ur

na

lP

re

contributions to the discussion.

27

Journal Pre-proof 8. References

Arnold, L.J., Bailey, R.M., Tucker, G.E., 2007. Statistical treatment of fluvial dose distributions from southern Colorado arroyo deposits. Quaternary Geochronology 2, 162-167. Arnold, L.J., Demuro, M., Navazo, M., Benito-Calvo, A., Pérez-González, A., 2012. OSL dating of the Middle Palaeolithic Hotel California site, Sierra de Atapuerca, north-central Spain. Boreas 42, 285305. Arnold, L.J., Roberts, R.G., Galbraith, R.F., DeLong, S.B., 2009. A revised burial dose estimation

of

procedure for optical dating of young and modern-age sediments. Quaternary Geochronology 4, 306-

ro

325.

Arnold, L.J., Roberts, R.G., MacPhee, R.D.E., Willerslev, E., Tikhonov, A.N., Brock, F., 2008.

-p

Optical dating of perennially frozen deposits associated with preserved ancient plant and animal DNA in

re

north-central Siberia. Quaternary Geochronology 3, 114-136.

Ayliffe, L.K., Marianelli, P.C., Moriarty, K.C., Wells, R.T., McCulloch, M.T., Mortimer, G.E.,

na

relative aridity. Geology 26, 147-150.

lP

Hellstrom, J.C., 1998. 500 ka precipitation record from southeastern Australia: Evidence for interglacial

Baczkowski, A.J., 2000. The broken-stick model for species abundances: An initial investigation.

ur

Internal Report STAT 00/10, University of Leeds, UK, 2-25. Bailey, R.M., Arnold, L.J., 2006. Statistical modelling of single grain quartz De distributions and an

Jo

assessment of procedures for estimating burial dose. Quaternary Science Reviews 25, 2475-2502. Ballhorn, U., Siegert, F., Mason, M., Limin, S., 2009. Derivation of burn scar depths and estimation of carbon emissions with LIDAR in Indonesian peatlands. Proceedings of the National Academy of Sciences 106, 21213. Barr, C., Tibby, J., Leng, M.J., Tyler, J.J., Henderson, A.C.G., Overpeck, J.T., Simpson, G.L., Cole, J.E., Phipps, S.J., Marshall, J.C., McGregor, G.B., Hua, Q., McRobie, F.H., 2019. Holocene El Niño– Southern Oscillation variability reflected in subtropical Australian precipitation. Scientific Reports 9, 1627.

28

Journal Pre-proof Barr, C., Tibby, J., Moss, P.T., Halverson, G.P., Marshall, J.C., McGregor, G.B., Stirling, E., 2017. A 25,000-year record of environmental change from Welsby Lagoon, North Stradbroke Island, in the Australian subtropics. Quaternary International 449, 106-118. Bateman, M.D., Boulter, C.H., Carr, A.S., Frederick, C.D., Peter, D., Wilder, M., 2007. Preserving the palaeoenvironmental record in Drylands: Bioturbation and its significance for luminescence-derived chronologies. Sedimentary Geology 195, 5-19. Beck, K.K., Fletcher, M.-S., Gadd, P.S., Heijnis, H., Jacobsen, G.E., 2017. An early onset of ENSO influence in the extra-tropics of the southwest Pacific inferred from a 14,600 year high resolution multi-

of

proxy record from Paddy's Lake, northwest Tasmania. Quaternary Science Reviews 157, 164-175.

ro

Berger, A., Crucifix, M., Hodell, D.A., Mangili, C., McManus, J.F., Otto-Bliesner, B., Pol, K., Raynaud, D., Skinner, L.C., Tzedakis, P.C., Wolff, E.W., Yin, Q.Z., Abe-Ouchi, A., Barbante, C., Brovkin, V.,

-p

Cacho, I., Capron, E., Ferretti, P., Ganopolski, A., Grimalt, J.O., Hönisch, B., Kawamura, K., Landais, A.,

re

Margari, V., Martrat, B., Masson-Delmotte, V., Mokeddem, Z., Parrenin, F., Prokopenko, A.A., Rashid, H., Schulz, M., Riveiros, N.V., 2015. Interglacials of the last 800,000 years. Reviews of Geophysics 54, 162-

lP

219.

Berntsson, A., Rosqvist, G.C., Velle, G., 2014. Late-Holocene temperature and precipitation

na

changes in Vindelfjällen, mid-western Swedish Lapland, inferred from chironomid and geochemical data. Holocene 24, 78-92.

ur

Bostock, H.C., Opdyke, B.N., Gagan, M.K., Kiss, A.E., Fifield, L.K., 2006. Glacial/interglacial

Jo

changes in the East Australian current. Climate Dynamics 26, 645-659. Bouchard, F., Francus, P., Pienitz, R., Laurion, I., 2011. Sedimentology and geochemistry of thermokarst ponds in discontinuous permafrost, subarctic Quebec, Canada. Journal of Geophysical Research: Biogeosciences 116. Bowler, J.M., Duller, G.A.T., Perret, N., Prescott, J.R., Wyrwoll, K.-H., 1998. Hydrologic changes in monsoonal climates of the last glacial cycle: stratigraphy and luminescence dating of Lake Woods, N.T. Australia. Paleoclimates 3, 179-207. Bowler, J.M., Wyrwoll, K.-H., Lu, Y., 2001. Variations of the northwest Australia summer monsoon over the last 300,000 years: The paleohydrological record of the Gregory (Mulan) Lakes system. Quaternary International 82-85, 63-80.

29

Journal Pre-proof Boyle, J.F., Chiverrell, R.C., Schillereff, D., 2015. Approaches to Water Content Correction and Calibration for μXRF Core Scanning: Comparing X-ray Scattering with Simple Regression of Elemental Concentrations, in: Croudace, I.W., Rothwell, R.G. (Eds.), Micro-XRF Studies of Sediment Cores, Developments in Paleoenvironmental Research. Springer Science and Business Media, Dordrecht. Brill, D., Klasen, N., Jankaew, K., Brückner, H., Kelletat, D., Scheffers, A., Scheffers, S., 2012. Local inundation distances and regional tsunami recurrence in the Indian Ocean inferred from luminescence dating of sandy deposits in Thailand. Natural Hazards and Earth System Sciences 12, 2177-2192. Brock, F., Higham, T., Ditchfield, P., Ramsey, C.B., 2010. Current Pretreatment Methods for AMS

of

Radiocarbon Dating at the Oxford Radiocarbon Accelerator Unit (ORAU). Radiocarbon 52, 103-112.

ro

Brock, F., Lee, S., Housley, R.A., Ramsey, B., 2011. Variation in the radiocarbon age of different fractions of peat: A case study from Ahrenshöft, northern Germany. Quaternary Geochronology 6, 550–

-p

555.

re

Bronk Ramsey, C., 2009. Bayesian Analysis of Radiocarbon Dates. Radiocarbon 51, 337-360.

lP

Bronk Ramsey, C., Lee, S., 2013. Recent and Planned Developments of the Program OxCal. Radiocarbon 55, 720-730.

na

Brumsack, H.-J., 2006. The trace metal content of recent organic carbon-rich sediments: Implications for Cretaceous black shale formation. Palaeogeography, Palaeoclimatology, Palaeoecology

ur

232, 344-361.

Burrows, M.A., Heijnis, H., Gadd, P., Haberle, S.G., 2016. A new late Quaternary

Jo

palaeohydrological record from the humid tropics of northeastern Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 451, 164-182. Cadd, H., Tibby, J., Barr, C., Tyler, J., Unger, L., Leng, M., Marshall, J., McGregor, G., Lewis, R., Arnold, L., Lewis, T., Baldock, J., 2018. Development of a southern hemisphere subtropical wetland (Welsby Lagoon, south-east Queensland, Australia) through the last glacial cycle. Quaternary Science Reviews 202, 53-65. Clark, M., Hounslow, M., 2009. CPLSlot Version 2.4b, Monash University, Victoria, Australia. Cohen, T.J., Jansen, J.D., Gliganic, L.A., Larsen, J.R., Nanson, G.C., May, J.-H., Jones, B.G., Price, D.M., 2015. Hydrological transformation coincided with megafaunal extinction in central Australia. Geology 43, 195-199. 30

Journal Pre-proof Colhoun, E.A., Pola, J.S., Barton, C.E., Heijnis, H., 1999. Late Pleistocene vegetation and climate history of Lake Selina, western Tasmania. Quaternary International 57–58, 5–23. Croudace, I.W., Rindby, A., Rothwell, R.G., 2006. ITRAX: description and evaluation of a new multi-function X-ray core scanner. Geological Society, London, Special Publications 267, 51. Croudace, I.W., Rothwell, R.G., 2015. Micro-XRF Studies of Sediment Cores: Applications of a non-destructive tool for the environmental sciences. Springer. Cuven, S., Francus, P., Lamoureux, S., 2010. Estimation of grain-size variability with micro X-ray

of

fluorescence in laminated lacustrine sediments, cape bounty, Canadian High Arctic. Journal of Paleoliminology 44, 803–817.

ro

Davies, G.M., Gray, A., Rein, G., Legg, C.J., 2013. Peat consumption and carbon loss due to

-p

smouldering wildfire in a temperate peatland. Forest Ecology and Management 308, 169-177. De Deckker, P., Arnold, L.J., van der Kaars, S., Bayon, G., Stuut, J.-B.W., Perner, K., Lopes dos

re

Santos, R., Uemura, R., Demuro, M., 2019. Marine Isotope Stage 4 in Australasia: A full glacial

lP

culminating 65,000 years ago – Global connections and implications for human dispersal. Quaternary Science Reviews 204, 187-207.

na

Demuro, M., Arnold, L.J., Froese, D.G., Roberts, R.G., 2013. OSL dating of loess deposits bracketing Sheep Creek tephra beds, northwest Canada: Dim and problematic single-grain OSL

ur

characteristics and their effect on multi-grain age estimates. Quaternary Geochronology 15, 67-87. Demuro, M., Arnold, L.J., Parés, J.M., Sala, R., 2015. Extended-range luminescence chronologies

Jo

suggest potentially complex bone accumulation histories at the Early-to-Middle Pleistocene palaeontological site of Huéscar-1 (Guadix-Baza basin, Spain). Quaternary International 389, 191-212. Donders, T.H., Wagner, F., Visscher, H., 2006. Late Pleistocene and Holocene subtropical vegetation dynamics recorded in perched lake deposits on Fraser Island, Queensland, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 241, 417-439. Edney, P.A., Kershaw, A.P., Deckker, P.D., 1990. A late Pleistocene and Holocene vegetation and environmental record from Lake Wangoom, Western Plains of Victoria, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 80, 325-343. Eggenberger, S., Gobet, E., Leeuwen, J., Schwörer, C., Knaap, W., Dobben, H., Vogel, H., Tinner, W., Rambeau, C., 2018. Millennial multi-proxy reconstruction of oasis dynamics in Jordan, by the Dead 31

Journal Pre-proof Sea. The Journal of Quaternary Plant Ecology, Palaeoclimate and Ancient Agriculture - Official Organ of the International Work Group for Palaeoethnobotany 27, 649-664. Elderfield, H., Ferretti, P., Greaves, M., Crowhurst, S., McCave, I.N., Hodell, D., Piotrowski, A.M., 2012. Evolution of Ocean Temperature and Ice Volume Through the Mid-Pleistocene Climate Transition. Science 337, 704-709. Falster, G., Tyler, J., Grant, K., Tibby, J., Turney, C., Löhr, S., Jacobsen, G., Kershaw, A.P., 2018. Millennial-scale variability in south-east Australian hydroclimate between 30,000 and 10,000 years ago.

of

Quaternary Science Reviews 192, 106-122. Field, E., Tyler, J., Gadd, P.S., Moss, P., McGowan, H., Marx, S., 2018. Coherent patterns of

ro

environmental change at multiple organic spring sites in northwest Australia: Evidence of IndonesianAustralian summer monsoon variability over the last 14,500 years. Quaternary Science Reviews 196,

-p

193-216.

re

Foerster, V., Deocampo, D.M., Asrat, A., Gunter, C., Junginger, A., Kramer, K.H., Stroncik, N.A., Trauth, M.H., 2018. Towards an understanding of climate proxy formation in the Chew Bahir basin,

lP

southern Ethiopian Rift. Palaeogeography, Palaeoclimatology, Palaeoecology 501, 111. Fu, X., Cohen, T.J., Arnold, L.J., 2017. Extending the record of lacustrine phases beyond the last

ur

162, 88-110.

na

interglacial for Lake Eyre in central Australia using luminescence dating. Quaternary Science Reviews

Grant, K.M., Rohling, E.J., Westerhold, T., Zabel, M., Heslop, D., Konijnendijk, T., Lourens, L.,

Jo

2017. A 3 million year index for North African humidity/aridity and the implication of potential panAfrican Humid periods. Quaternary Science Reviews 171, 100-118. Guyard, H., Chapron, E., St-Onge, G., Anselmetti, F.S., Arnaud, F., Magand, O., Francus, P., Mélières, M.-A., 2007. High-altitude varve records of abrupt environmental changes and mining activity over the last 4000 years in the Western French Alps (Lake Bramant, Grandes Rousses Massif). Quaternary Science Reviews 26, 2644-2660. Hahn, A., Kliem, P., Oehlerich, M., Ohlendorf, C., Zolitschka, B., 2014. Elemental composition of the Laguna Potrok Aike sediment sequence reveals paleoclimatic changes over the past 51 ka in southern Patagonia, Argentina. Journal of Paleolimnology 52, 349-366.

32

Journal Pre-proof Harle, K.J., Heijnis, H., Chisari, R., Kershaw, A.P., Zoppi, U., Jacobsen, G., 2002. A chronology for the long pollen record from Lake Wangoom, western Victoria (Australia) as derived from uranium/thorium disequilibrium dating. Journal of Quaternary Science 17, 707-720. Harrison, S.P., Bartlein, P., 2012. Records from the Past, Lessons for the Future: What the Palaeorecord Implies about Mechanisms of Global Change, in: Henderson-Sellers, A., McGuffie, K. (Eds.), The Future of the World's Climate. Elsevier Science, pp. 403-436. Heiri, O., Lotter, A., Lemcke, G., 2001. Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. Journal of Paleolimnology

of

25, 101-110.

ro

Hesse, P.P., McTainsh, G.H., 2003. Australian dust deposits: modern processes and the

-p

Quaternary record. Quaternary Science Reviews 22, 2007-2035.

Hofmann, H., Newborn, D., Cartwright, I., Cendón, D.I., Raiber, M., 2019. Groundwater mean

re

residence time of a sub-tropical barrier sand island. Hydrology and Earth System Science Discussions 1-

lP

32.

Hogg, A.G., Hua, Q., Blackwell, P.G., Niu, M., Buck, C.E., Heaton, T.J., Guilderson, T.P., Zimmerman, S.R.H., Palmer, J.G., Turney, C.S.M., Reimer, P.J., Reimer, R.W., 2013. Shcal13 Southern

na

Hemisphere calibration, 0-50,000 years cal BP. Radiocarbon 55, 1889-1903.

Statistical Computing.

ur

Ihaka, R., Gentleman, R., 2016. R statistical computing program, 3.3.1 ed. The R Foundation for

Jo

Jouve, G., Francus, P., Lamoureux, S., Provencher-Nolet, L., Hahn, A., Haberzettl, T., Fortin, D., Nuttin, L., 2013. Microsedimentological characterization using image analysis and μ-XRF as indicators of sedimentary processes and climate changes during Lateglacial at Laguna Potrok Aike, Santa Cruz, Argentina. Quaternary Science Reviews 71, 191-204. Kawagata, S., 2001. Tasman Front shifts and associated paleoceanographic changes during the last 250,000 years: foraminiferal evidence from the Lord Howe Rise. Marine Micropaleontology 41, 167191. Kemp, C.W., Tibby, J., Arnold, L.J., Barr, C., 2019. Australian hydroclimate during Marine Isotope Stage 3: A synthesis and review. Quaternary Science Reviews 204, 94-104.

33

Journal Pre-proof Kershaw, P., Moss, P., Van Der Kaars, S., 2003. Causes and consequences of long‐term climatic variability on the Australian continent. Freshwater Biology 48, 1274-1283. Kershaw, A.P., Bretherton, S.C., van der Kaars, S., 2007a. A complete pollen record of the last 230 ka from Lynch's Crater, north-eastern Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 251, 23-45. Kershaw, A.P., McKenzie, G.M., Porch, N., Roberts, R.G., Brown, J., Heijnis, H., Orr, M.L., Jacobsen, G., Newall, P.R., 2007b. A high-resolution record of vegetation and climate through the last glacial cycles from Caledonia Fen, southeastern highlands of Australia. Journal of Quaternary Science 22,

of

481-500.

ro

Kienel, U., Kirillin, G., Brademann, B., Plessen, B., Lampe, R., Brauer, A., 2017. Effects of spring warming and mixing duration on diatom deposition in deep Tiefer See, NE Germany. Journal of

-p

Paleolimnology 57, 37-49.

lP

Statistical Software; Vol 1, Issue 3 (2014).

re

Killick, R., Eckley, I.A., 2014. Changepoint: An R Package for Changepoint Analysis. Journal of

Köhler, P., Nehrbass-Ahles, C., Schmitt, J., Stocker, T.F., Fischer, H., 2017. Continuous record of the atmospheric greenhouse gas carbon dioxide (CO2), raw data, In supplement to: Köhler, P et al.

na

(2017): A 156 kyr smoothed history of the atmospheric greenhouse gases CO2, CH4, and N2O and their

PANGAEA.

ur

radiative forcing. Earth System Science Data, 9(1), 363-387, https://doi.org/10.5194/essd-9-363-2017.

Jo

Kylander, M.E., Ampe, L., Wohlfarth, B., Veres, D., 2011. High-resolution X-ray fluorescence core scanning analysis of Les Echets (France) sedimentary sequence: new insights from chemical proxies. Journal of Quaternary Science 26, 109-117. Leach, L.M., 2011. Hydrology and physical setting of North Stradbroke Island. Proceedings of the Royal Society of Queensland 117, 21-46. Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene-Pleistocene stack of 57 globally distributed benthic δ18O records. Paleoceanography 20. Lo, L., Chang, S.-P., Wei, K.-Y., Lee, S.-Y., Ou, T.-H., Chen, Y.-C., Chuang, C.-K., Mii, H.-S., Burr, G.S., Chen, M.-T., Tung, Y.-H., Tsai, M.-C., Hodell, D.A., Shen, C.-C., 2017. Nonlinear climatic sensitivity to greenhouse gases over past 4 glacial/interglacial cycles. Scientific Reports 7, 4626. 34

Journal Pre-proof Longmore, M.E., 1997. The mid-Holocene "Dry" Anomaly on the Mid-Eastern Coast of Australia: Calibration of Palaeowater Depth as a Surrogate for Effective Precipitation using Sedimentary Loss on Ignition in the Perched Lake Sediments of Fraser Island, Queensland. Palaeoclimates 4:46023. Löwemark, L., Chen, H.F., Yang, T.N., Kylander, M., Yu, E.F., Hsu, Y.W., Lee, T.Q., Song, S.R., Jarvis, S., 2011. Normalizing XRF-scanner data: a cautionary note on the interpretation of high-resolution records from organic-rich lakes. Journal of Asian Earth Sciences 40, 1250–1256. Lukenbach, M.C., Hokanson, K.J., Moore, P.A., Devito, K.J., Kettridge, N., Thompson, D.K., Wotton, B.M., Petrone, R.M., Waddington, J.M., 2015. Hydrological controls on deep burning in a

of

northern forested peatland. Hydrological Processes 29, 4114-4124.

ro

Mackenzie, L., Heijnis, H., Gadd, P., Moss, P., Shulmeister, J., 2017. Geochemical investigation of the South Wellesley Island wetlands: Insight into wetland development during the Holocene in tropical

-p

northern Australia. The Holocene 27, 566-578.

re

Maroulis, J.C., Nanson, G.C., Price, D.M., Pietsch, T., 2007. Aeolian–fluvial interaction and climate change: source-bordering dune development over the past ∼100 ka on Cooper Creek, central

lP

Australia. Quaternary Science Reviews 26, 386-404.

Marshall, M.H., Lamb, H.F., Huws, D., Davies, S.J., Bates, R., Bloemendal, J., Boyle, J., Leng, M.J.,

na

Umer, M., Bryant, C., 2011. Late Pleistocene and Holocene drought events at Lake Tana, the source of

ur

the Blue Nile. Global and Planetary Change 78, 147-161. Martin, L., Mooney, S., Goff, J., 2014. Coastal wetlands reveal a non-synchronous island

Jo

response to sea-level change and a palaeostorm record from 5.5 kyr to present. The Holocene 24, 569580.

Martinez, J.I., DeDeckker, P., Barrows, T.T., 2002. Palaeoceanography of the western Pacific warm pool during the last glacial maximum: long-term climatic monitoring of the maritime continent., in: Kershaw, P., Bruno, D., Tapper, N., Penny, D., Brown, J. (Eds.), Bridging Wallace’s line. Adv GeoEcol, pp. 147-172. McGowan, H.A., Petherick, L.M., Kamber, B.S., 2008. Aeolian sedimentation and climate variability during the late Quaternary in southeast Queensland, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 265, 171-181.

35

Journal Pre-proof Miller, G.H., Fogel, M.L., Magee, J.W., Gagan, M.K., 2016. Disentangling the impacts of climate and human colonization on the flora and fauna of the Australian arid zone over the past 100 ka using stable isotopes in avian eggshell. Quaternary Science Reviews 151, 27-57. Moss, P.T., Kershaw, A.P., 2007. A late Quaternary marine palynological record (oxygen isotope stages 1 to 7) for the humid tropics of northeastern Australia based on ODP Site 820. Palaeogeography, Palaeoclimatology, Palaeoecology 251, 4-22. Moss, P.T., Tibby, J., Petherick, L., McGowan, H., Barr, C., 2013. Late Quaternary vegetation history of North Stradbroke Island, Queensland, eastern Australia. Quaternary Science Reviews 74, 257-

of

272.

ro

Nanson, G.C., Price, D.M., Jones, B.G., Maroulis, J.C., Coleman, M., Bowman, H., Cohen, T.J., Pietsch, T.J., Larsen, J.R., 2008. Alluvial evidence for major climate and flow regime changes during the

-p

middle and late Quaternary in eastern central Australia. Geomorphology 101, 109-129.

re

Neal, R., Stock, E., 1986. Pleistocene occupation in the south-east Queensland coastal region.

lP

Nature 323, 618-621.

Nilsson, M., Klarqvist, M., Bohlin, E., Possnert, G., 2001. Variation in 14C age of macrofossils and

na

different fractions of minute peat samples dated by AMS. The Holocene 11, 579–586. Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin, P.R., R. B.

ur

O'Hara, Simpson, G.L., Solymos, P., Stevens, M.H.H., Szoecs, E., Wagner, H., 2018. Community Ecology Package, 2.5-2 ed, pp. Ordination methods, diversity analysis and other functions for community and

Jo

vegetation ecologists.

Parrenin, F., Masson-Delmotte, V., Köhler, P., Raynaud, D., Paillard, D., Schwander, J., Barbante, C., Landais, A., Wegner, A., Jouzel, J., 2013. Antarctic Temperature Stack (ATS) from five different ice cores (EDC, Vostok, Dome Fuji, TALDICE, and EDML), In supplement to: Parrenin, F et al. (2013): Synchronous change of atmospheric CO2 and Antarctic temperature during the last deglacial warming. Science, 339 (6123), 1060-1063, https://doi.org/10.1126/science.1226368. PANGAEA. Patton, N.R., Ellerton, D., Shulmeister, J., 2019. High resolution remapping of the coastal dune fields of south east Queensland, Australia: a morphometric approach. Journal of Maps 15, 578-589. Pelejero, C., Calvo, E., Barrows, T.T., Logan, G.A., De Deckker, P., 2006. South Tasman Sea alkenone palaeothermometry over the last four glacial/interglacial cycles. Marine Geology 230, 73-86. 36

Journal Pre-proof Petherick, L., Bostock, H., Cohen, T.J., Fitzsimmons, K., Tibby, J., Fletcher, M.-S., Moss, P., Reeves, J., Mooney, S., Barrows, T., Kemp, J., Jansen, J., Nanson, G., Dosseto, A., 2013. Climatic records over the past 30 ka from temperate Australia – a synthesis from the Oz-INTIMATE workgroup. Quaternary Science Reviews 74, 58-77. Petherick, L., McGowan, H., Moss, P., 2008. Climate variability during the Last Glacial Maximum in eastern Australia: evidence of two stadials? Journal of Quaternary Science Review 23, 787-802. Petherick, L.M., McGowan, H.A., Kamber, B.S., 2009. Reconstructing transport pathways for late Quaternary dust from eastern Australia using the composition of trace elements of long travelled dusts.

of

Geomorphology 105, 67-79.

-p

subtropical Australia. Quaternary International 432, 1-12.

ro

Petherick, L.M., Moss, P.T., McGowan, H.A., 2017. An extended Last Glacial Maximum in

Pleskot, K., Tjallingii, R., Makohonienko, M., Nowaczyk, N., Szczuciński, W., 2018. Holocene

re

paleohydrological reconstruction of Lake Strzeszyńskie (western Poland) and its implications for the

lP

central European climatic transition zone. Journal of Paleolimnology 59, 443-459. Profe, J., Neumann, L., Novothny, Á., Barta, G., Rolf, C., Frechen, M., Ohlendorf, C., Zolitschka, B., 2018. Paleoenvironmental conditions and sedimentation dynamics in Central Europe inferred from

na

geochemical data of the loess-paleosol sequence at Süttő (Hungary). Quaternary Science Reviews 196,

ur

21-37.

Rees, A.B.H., Cwynar, L.C., Fletcher, M.-S., 2015. Southern Westerly Winds submit to the ENSO

126, 254-263.

Jo

regime: A multiproxy paleohydrology record from Lake Dobson, Tasmania. Quaternary Science Reviews

Reeves, J.M., Barrows, T.T., Cohen, T.J., Kiem, A.S., Bostock, H.C., Fitzsimmons, K.E., Jansen, J.D., Kemp, J., Krause, C., Petherick, L., Phipps, S.J., Members, O.-I., 2013. Climate variability over the last 35,000 years recorded in marine and terrestrial archives in the Australian region: an OZ-INTIMATE compilation. Quaternary Science Reviews 74, 21-34. Roberts, H.M., Bryant, C.L., Huws, D.G., Lamb, H.F., 2018. Generating long chronologies for lacustrine sediments using luminescence dating: a 250,000 year record from Lake Tana, Ethiopia. Quaternary Science Reviews 202, 66-77.

37

Journal Pre-proof Rohling, E.J., Grant, K., Bolshaw, M., Roberts, A.P., Siddall, M., Hemleben, C., Kucera, M., 2009. Antarctic temperature and global sea level closely coupled over the past five glacial cycles. Nature Geoscience 2, 500. Sáez, A., Valero-Garcés, B.L., Giralt, S., Moreno, A., Bao, R., Pueyo, J.J., Hernández, A., Casas, D., 2009. Glacial to Holocene climate changes in the SE Pacific. The Raraku Lake sedimentary record (Easter Island, 27°S). Quaternary Science Reviews 28, 2743-2759. Smith, S.M., Newman, S., Garrett, P.B., Leeds, J.A., 2001. Differential Effects of Surface and Peat Fire on Soil Constituents in a Degraded Wetland of the Northern Florida Everglades. Journal of

of

Environmental Quality 30, 1998-2005.

ro

Stephens, T., Augustinus, P., Rip, B., Gadd, P., Zawadski, A., 2018. Managing land-use effects on Northland dune lakes: lessons from the past. New Zealand Journal of Marine and Freshwater Research,

-p

1-21.

re

Striberger, J., Björck, S., Ingólfsson, Ó., Kjær, K.H., Snowball, I.A.N., Uvo, C.B., 2010. Climate variability and glacial processes in eastern Iceland during the past 700 years based on varved lake

lP

sediments. Boreas 40, 28-45.

Tachikawa, K., Timmermann, A., Vidal, L., Sonzogni, C., Timm, O.E., 2014. CO2 radiative forcing

na

and Intertropical Convergence Zone influences on western Pacific warm pool climate over the past 400

ur

ka. Quaternary Science Reviews 86, 24-34.

Team-RStudio, 2015. RStudio: Integrated Development for R. RStudio, Inc, Boston, MA.

Jo

Tibby, J., Barr, C., Marshall, J.C., Mcgregor, G.B., Moss, P.T., Arnold, L.J., Page, T.J., Questiaux, D., Olley, J., Kemp, J., Spooner, N., Petherick, L., Penny, D., Mooney, S., Moss, E., 2017. Persistence of wetlands on North Stradbroke Island (south‐east Queensland, Australia) during the last glacial cycle: implications for Quaternary science and biogeography. Journal of Quaternary Science 32, 770-781. Turetsky, M.R., Manning, S.W., Wieder, R.K., 2004. Dating recent peat deposits. Wetlands 24, 324-356. Turner, J.N., Holmes, N., Davis, S.R., Leng, M.J., Langdon, C., Scaife, R.G., 2015. A multiproxy (micro‐XRF, pollen, chironomid and stable isotope) lake sediment record for the Lateglacial to Holocene transition from Thomastown Bog, Ireland. Journal of Quaternary Science 30, 514-528.

38

Journal Pre-proof Turner, T.E., Swindles, G., Roucoux, K., 2014. Late Holocene ecohydrological and carbon dynamics of a UK raised bog: impact of human activity and climate change. Quarternary Science Reviews 84, 65-85. Turney, C.S.M., Jones, R.T., 2010. Does the Agulhas Current amplify global temperatures during super‐interglacials? Journal of Quaternary Science 25, 839-843. Unkel, I., Fernandez, M., Björck, S., Ljung, K., Wohlfarth, B., 2010. Records of environmental changes during the Holocene from Isla de los Estados (54.4°S), southeastern Tierra del Fuego. Global

of

Planet Change 74, 99–113. Van der Bilt, W.G.M., Bakke, J., Vasskog, K., D'Andrea, W.J., Bradley, R.S., Ólafsdóttir, S., 2015.

ro

Reconstruction of glacier variability from lake sediments reveals dynamic Holocene climate in Svalbard.

-p

Quaternary Science Reviews 126, 201-218.

Van der Kaars, S., Miller, G.H., Turney, C.S.M., Cook, E.J., Nürnberg, D., Schönfeld, J., Kershaw,

re

A.P., Lehman, S.J., 2017. Humans rather than climate the primary cause of Pleistocene megafaunal

lP

extinction in Australia. Nature Communications 8, 1-7.

Vegas-Vilarrúbia, T., Corella, J.P., Pérez-Zanón, N., Buchaca, T., Trapote, M.C., López, P., Sigró, J., Rull, V., 2018. Historical shifts in oxygenation regime as recorded in the laminated sediments of lake

na

Montcortès (Central Pyrenees) support hypoxia as a continental-scale phenomenon. Science of The

ur

Total Environment 612, 1577-1592.

Walker, J., Lees, B., Olley, J., Thompson, C., 2018. Dating the Cooloola coastal dunes of South-

Jo

Eastern Queensland, Australia. Marine Geology 398, 73-85. Weltje, G.J., Bloemsma, M.R., Tjallingii, R., Heslop, D., Röhl, U., Croudace, I.W., 2015. Prediction of Geochemical Composition from XRF Core Scanner Data: A New Multivariate Approach Including Automatic Selection of Calibration Samples and Quantification of Uncertainties, in: Croudace, I.W., Rothwell, R.G. (Eds.), Micro-XRF Studies of Sediment Cores: Applications of a non-destructive tool for the environmental sciences. Springer Netherlands, Dordrecht, pp. 507-534. Weltje, G.J., Tjallingii, R., 2008. Calibration of XRF core scanners for quantitative geochemical logging of sediment cores: Theory and application. Earth and Planetary Science Letters 274, 423-438.

39

Journal Pre-proof Woltering, M., Atahan, P., Grice, K., Heijnis, H., Taffs, K., Dodson, J., 2014. Glacial and Holocene terrestrial temperature variability in subtropical east Australia: branched GDGT distributions in a sediment core from Lake McKenzie. Quaternary Research 82, 132–145. Woodward, C.A., Slee, A., Gadd, P., Zawadzki, A., Hamze, H., Parmar, A., Zahra, D., 2018. The role of earthquakes and climate in the formation of diamictic sediments in a New Zealand mountain lake. Quaternary International 470, 130-147. Woodward, C.A., Gadd, P.S., 2019. The potential power and pitfalls of using the X-ray fluorescence molybdenum incoherent:coherent scattering ratio as a proxy for sediment organic content.

of

Quaternary International 514, 30-43.

ro

Wright, H.E.J., 1967. A square-rod piston sampler for lake sediments. The Journal of

-p

Sedimentary Petrology 37, 975–976.

Zhang, S., Li, T., Chang, F., Yu, Z., Xiong, Z., Wang, H., 2017. Correspondence between the ENSO-

re

like state and glacial-interglacial condition during the past 360 kyr. Chinese Journal of Oceanology and

Jo

ur

na

lP

Limnology 35, 1018-1031.

40

Journal Pre-proof Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal

of

relationships that could have appeared to influence the work reported in this paper.

Jo

ur

na

lP

re

-p

ro

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

41

Journal Pre-proof Climates of the last three interglacials in subtropical eastern Australia inferred from wetland sediment geochemistry: Figures and tables

Kemp, C.W.1, Tibby, J.1, Arnold, L.J.2, Barr, C.1, Gadd, P.S.3, Marshall, J.C.4,5, McGregor, G.B.4, Jacobsen, G.E.3

North Terrace Campus, Adelaide, 5005, SA, Australia

of

6. Geography, Environment and Population and Sprigg Geobiology Centre, University of Adelaide,

ro

7. School of Physical Science, Environment Institute, Sprigg Geobiology Centre and Institute for

-p

Photonics and Advanced Sensing, University of Adelaide, North Terrace Campus, Adelaide, 5005, SA, Australia

re

8. Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, New South Wales 2232, Australia

lP

9. Queensland Department of Environment and Science, Dutton Park, Queensland, Australia

Jo

ur

na

10. Australian Rivers Institute, Griffith University, Nathan, 4111, Queensland, Australia

42

Jo

ur

na

lP

re

-p

ro

of

Journal Pre-proof

Figure 1: A: Modified Köppen climate zones (BOM, 2005), location of Australian interglacial climate records and the approximate current positions of the west Pacific warm pool (WPWP, mean annual SST >28oC) boundary (pink shading; (Gagan et al., 2000)), the East Australian Current (EAC) and the Tasman Front (TF) (Bostock et al., 2006). B: North Stradbroke Island, with the four climate/archaeological record sites (black circles) and local towns (white circles), mentioned in the text. C: Fern Gully Lagoon combined topographic map and satellite image, indicating the coring location, locations of modern geochemistry samples, outflows and the height of surrounding dunes (m ASL).

43

Jo

ur

na

lP

re

-p

ro

of

Journal Pre-proof

Figure 2: Relative position of FG15-2 to FG15-1 using the seven elements exhibiting the highest counts per second

44

Journal Pre-proof

Element Zircon (Zr)

quantified by XRF pure element (ppm)

Titanium (Ti) Potassiu m (K) Silicon (Si) Bromine (Br)

TiO2

K2O

SiO2 pure element (ppm)

Calcium (Ca) Sulfur (S)

CaO

SO3

Iron (Fe)

Fe2O3

Gallium

pure

(Ga)

element (ppm) Aluminu

m (Al) Magnesiu m (Mg) Phosphor us (P)

Material

Al2O3

MgO

P2O5

Total cts/s normalised

Centered log

Al

Raw cts/s

ratio

Total scatter

normalised

normalised

Incoherent scatter normalised

r2 = 0.93, p <

r2 = 0.65, p <

r2 = 0.92, p <

r2 = 0.83, p <

r2 = 0.92, p <

r2 = 0.92, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.88, p <

r2 = 0.66, p <

r2 = 0.90, p <

r2 = 0.80, p <

r2 = 0.90, p <

r2 = 0.89, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.83, p <

r2 = 0.79, p <

r2 = 0.85, p <

r2 = 0.73, p <

r2 = 0.85, p <

r2 = 0.84, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.80, p <

r2 = 0.57, p <

r2 = 0.78, p <

r2 = 0.65, p <

r2 = 0.77, p <

r2 = 0.76, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.71, p <

r2 = 0.84, p <

r2 = 0.44, p <

r2 = 0.61, p <

r2 = 0.44, p <

r2 = 0.42, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.74, p <

r2 = 0.78, p <

r2 = 0.42, p <

r2 = 0.54, p <

r2 = 0.42, p <

r2 = 0.43, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.73, p <

r2 = 0.70, p <

r2 = 0.30, p <

r2 = 0.33, p <

r2 = 0.28, p <

r2 = 0.26, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.53, p <

r2 = 0.55, p <

r2 = 0.50, p <

r2 = 0.44, p <

r2 = 0.50, p <

r2 = 0.49, p <

0.05

0.05

0.05

0.05

0.05

0.05

r2 = 0.11, p =

r2 = 0.19, p =

r2 = 0.03, p =

r2 = 0.20, p <

r2 = 0.01, p =

r2 = 0.01, p =

0.16

0.05

0.48

0.05

0.69

0.72

r2 = 0.09, p =

r2 = 0.09, p =

r2 = 0.62, p <

r2 = 0.67, p <

r2 = 0.69, p <

0.19

0.19

0.05

0.05

0.05

r2 = 0.05, p =

r2 = 0.85, p <

r2 < 0.01, p =

r2 = 0.08, p =

r2 = 0.02, p =

r2 = 0.02, p =

0.33

0.05

0.79

0.22

0.60

0.55

r2 = 0.02, p =

r2 = 0.46, p <

r2 = 0.11, p =

r2 < 0.01, p =

r2 = 0.12, p =

r2 = 0.13, p =

0.56

0.05

0.14

0.76

0.13

0.12

P l a

n r u

o J

ro

-p

re

N/A

f o

Table 1: Regressions of known quantities of oxides and elements via WD-XRF versus corrected µXRF data with methods ranked first in order of coefficient of determination then by number of p values < 0.05 (in bold). n = 20 for all r2 values. Titanium, silicon and zircon normalisation had r2 < 0.5, p > 0.05 in all cases and are not shown. cts/s: µXRF counts per second.

45

Jo

ur

na

lP

re

-p

ro

of

Journal Pre-proof

46

Journal Pre-proof Figure 3: Fern Gully Lagoon sediment stratigraphy, visible wavelength photography (brightened to show colour variation), sediment bulk density, water content, magnetic susceptibility, X-radiograph derived relative density, inorganic content and µXRF derived SiO2 content. The four depositional boundaries used in the age model were derived

Jo

ur

na

lP

re

-p

ro

of

from silicon µXRF counts per second (cts/s, section 2.4 and supplementary material).

47

Journal Pre-proof

Calib 14

Sam ple ID

Dep

L ab ID

δ

th

Sample Type

pM 13

C

14

(cm)

14

onal C Age ( C yr

C (‰)

(ANSTO)

Conventi

(%)

BP)

rated C 95.4% probability range (cal. yr BP)

Fern Gully Core 1 (FG15-1)

FG15 RC-11

O ZU792

FG15 RC-12*

fragment

ZU793

RC-3*

ZU191

FG15 RC-5

O ZU193

FG15 RC-13

fragment O

ZU794 FG15

RC-14*

ZU795

RC-6*

O

FG15 RC-7*

5–175.5 O

ZU195 FG15

RC-8*

Reed stalk segment

O

RC-15**

Charcoal

224.

O

Terrestrial bark fragment

O

5–228.5 Charcoal

120

7–32682 165 ± 50

Mod

39,880 ± 330

0.71 ± 0.07

-

3163

ern–281

± 0.03

24.6 ± 0.1 227.

28,330 ±

0.70

-

1736 –1896

7 ± 0.55

26.7 ± 0.1 227.

2–19870

30

97.9

-

1942

1,925 ±

2.94

22.1 ± 0.3

5–225.5

ZU796 FG15

214.

2–17849

60

± 0.04

1739

16,290 ±

78.6

28.5 ± 0.1

5–215.5

ZU196 FG15

174.

14,480 ±

13.1

9 ± 0.27

1796 6–18366

60

6 ± 0.10

b

Mod

15,000 ± 60

9 ± 0.11

25.0

–7238 Modern

16.4

-

6952

ern

6 ± 0.11

25.0 ± 0.2

5–159.5

35

15.4

-

158.

6–13211

18 ± 0.61

22.6 ± 0.1

5–159.5

Charcoal

ZU194

134.

1303

6,225 ±

128.

-

158.

Reed stalk

segment

25.7 ± 0.1

655–

11,290 ± 40

7 ± 0.18

24.1 ± 0.1

5–135.5

Terrestrial bark

fragment O

FG15

Terrestrial leaf

–960

730

46.0

-

134.

5–135.5

-

803

790 ± 30

24.5

2 ± 0.13

a

114.

Terrestrial seed

ZU192

24.9

5–115.5

Jo ur

RC-4

Reed stalk

–79.5

30

2 ± 0.28

24.6 ± 0.1

78.5

1,040 ±

90.6

-

Reed stalk

segment O

-

26.2 ± 0.1 78.5

–79.5

segment O

FG15

–35.5 Terrestrial bark

O

FG15

34.5

fragment

3 ± 0.31

of

ZU190

Terrestrial leaf

26.0 ± 0.2

87.8

ro

RC-2

O

–35.5

-

re

FG15

nut

34.5

-p

ZU189

Terrestrial gum

lP

RC-1

O

na

FG15

6–44192 39,730 ±

810 0.35

4291

4234 3–44960

45,400 ±

4637

48

Journal Pre-proof RC-16*

ZU797 FG15

RC-9*

O

Charcoal

RC-10*

5–235.5 O

ZU198 FG15

RC-17*

Reed stalk

5–288.5 O

Reed stalk

26.4 ± 0.1

26.3

5–34803 44,110 ±

850 0.96

± 0.04

3394

4578 5–49430

37,300 ± 360

4117 8–42297

of

5–288.5

30,420 ±

0.41

a

4681 5–49875

240

± 0.04

c

48,200 ±

2.27

-

7–

750

± 0.07

287.

segment

0.25

23.4

1,500

± 0.02

a

287.

ZU798

ZU799

27.2 ± 0.1

5–275.5 Charcoal

± 0.06 -

274.

segment O

FG15

26.5 ± 0.3 234.

ZU197 FG15

RC-18*

5–228.5

Table 2: AMS Radiocarbon ages. * denotes an age excluded from the age-depth model due to material type, while **

ro

denotes an age eliminated from the final Bayesian model as it was identified as a major statistical outlier during initial modelling and prevented successful convergence. a δ13C values without associated uncertainty due to a limited

-p

number of determinations. b δ13C is assumed - measured value was not available. c Maximum value beyond calibration range. 14C ages have been calibrated using the SHCal13 curve (Hogg et al. 2013) in OxCal v4.2.4. The

re

calibrated age range shown is the 95.4% probability range (combining two or more potential calibration ranges,

lP

where they exist). All δ¹³C values relate solely to the graphite derived from the fraction that was used for the radiocarbon measurement and have been derived using EA-IRMS. Uncalibrated 14C ages have been corrected for

Jo ur

na

isotopic fractionation using their measured δ13C values and are quoted with their 1σ errors.

49

lP

re

-p

ro

of

Journal Pre-proof

Figure 4: Calibrated radiocarbon age probability distributions with calibrated 68.2% and

Jo ur

terrestrial source.

na

95.4% age ranges. * denotes an age excluded from the age-depth model due to material type. (T): a

50

Jo ur

na

lP

re

-p

ro

of

Journal Pre-proof

Figure 5: Example single-grain De distributions shown as radial plots and frequency histograms with ranked plots of De versus standard error. (a) Samples FG15 OSL-3 (813–832 cm depth), which was considered suitable for dating using the central age model (CAM). The grey bar on the radial plot is centred on the CAM De value used to derive the final burial dose of this sample; (b) Sample FG15 OSL-9 (326–335 cm depth), which was not considered suitable for dating because it contains multiple discrete dose populations, as identified by the finite mixture model (FMM). The two dose components identified by the FMM are shown by the light and dark grey bands in the radial plot. Individual De estimates are presented with their 1σ error ranges, which are derived from a random uncertainty term arising from photon counting statistics for each OSL measurement, an empirically determined instrument reproducibility uncertainty of 2.5% (following the approach 51

Journal Pre-proof outlined in Jacobs et al. (2006)) and a dose-response curve fitting uncertainty determined using 1000

Jo ur

na

lP

re

-p

ro

of

iterations of the Monte Carlo method implemented in Analyst (Duller, 2015).

52

Journal Pre-proof

Sa

Environmental dose rate (Gy/kyr)

W

mple

Equivalent dose (De)

b,c,d,e,f

data

D ater epth (c

Content

(cm)

(% dry

ore)

weight)

F G15 OSL18 (1)

1 71–180

F G15 OSL17 (1)

G15 OSL16 2 (2)

81–190

04–213

G15 OSL16 (1)

k

04–213

F G15 OSL-

3±0.005

42/361/4

86/344/4

1±0.003

0.027 ±0.001

0.0 3±0.005

o. of

dose rate

dose rate

rate

grains

0.

02±0.007

0.

0. 02±0.007

r P

03±0.003

l a

0.

11±0.01

0.

03±0.003

0. 03±0.003

03±0.003

11±0.01

12±0.01

9 6

1.3±3.6

(ka)

.60±0.8

7 7.7±9.2

1 4.3±0.6

AM-3

3 3.2±8.0

.08±0.7

M

1 28±15.3

1 .48±0.1

C AM

f,j

9

M

4

nal age

3

C

9

Fi

g

AM-3

2.2±10.7

f

M

3

5

0.

Model

AM

9.5±2.9

(Gy)

ge

5

1

0

e

AM-3

0.9±5.8

17

A

5

0.3±4.2

5

0.

0. 03±0.003

02

0.

0.

i

2

09±0.01

f o

on (σb) (%)

h

1

0.

12±0.01

N verdispersi

o r p

e

0.

n r u 02±0.007

0.03± 0.0009

otal dose

0.

o J

0.0006

3±0.006

osmic

02±0.007

0.02±

T

ternal

02±0.007

0.03±

0.0

C

0.

0.0009

0.0

3 36/363/4

0.03± 0.0007

3±0.005

3 86/344/4

ma dose rate

0.0

3

In

Gam

0.0

3

84 2

14–223

20/404/5

84 2

a dose rate

3

97 2

F

a

07 1

F

Bet

D

O

1 3.3±2.0

1 1.4±0.5

9 5.5±9.73

53

Journal Pre-proof 15 (1)

77 F

G15 OSL14 (1)

2 35–244

F G15 OSL13 (1)

G15 OSL12 (1)

63–272

73–282

G15 OSL11 (1)

83–292

F G15 OSL10 2 (2)

06–315

G15 OSL10 (1)

k

06–315

F G15 OSL-9 (1)

k

46/426/4

2±0.003

45/433/4

55

0.0 1±0.003

l a

n r u

0.007

o J

±0.0003

0.01±

0.0005

0.0 1±0.002

0. 02±0.007

0.0

3 80/450/4

0.02±

03±0.003

02±0.007

0. 02±0.007

0.02± 0.003

0.

0. 02±0.007

03±0.003

0.

03±0.003

0. 03±0.003

0. 03±0.003

3

0.

4.5±5.4

7

0.

31

0.

09±0.02

3

0. 08±0.009

8

0. 08±0.009

8 2

1

C

71±33.4

3.8±0.6

1 79±22.4

1 .65±0.1

M AM-4

1

1

M

1

1 37±15.4

4.8±0.5

AM-3

18.5±10.6

1

C

5

1 32±20.3

6.4±0.5

AM

3.3±6.1

1

C

3

1 35±16.3

4.7±0.9

AM

3.2±3.7

6

AM

2

6

2.7±0.9

AM

8.9±2.8

1

C

3

4.4±2.5

2

06±0.008

f o

8

0.

AM

4.0±5.0

1

C

3

o r p

e

1±0.01

3

3

1±0.01

r P 0.

3

09±0.009

0.

02±0.007

0.01

04±0.001

03±0.003

0.

0.003

0.

0.

02±0.007

0.03±

0.0

03±0.003

0.

0.008

4±0.007

0.

02±0.007

0.04±

0.0

5

0.

0.0007

2±0.004

5 45/433/4

0.03±

0.0

4

57 3

26–335

62/402/4

57 3

2±0.003

3

56 3

F

06/399/4

56 2

0.0

4

60 2

F

08/378/4 68

2

F

4

2 1.2±2.6

1 .59±0.1

2 0.0±2.8

54

Journal Pre-proof F G15 OSL-8 (1)

k

4 05–414

F G15 OSL-8 2 (2)

k

G15 OSL-5 (1)

05–414

55–674

G15 OSL-4 (2)

77–801

F G15 OSL-3 (1)

G15 OSL-2 (1)

13–832

G15 OSL-1 (2)

k

33–857

4 38/477/5

4 43/439/5

3 97/434/5 67

8 93–926

29/666/5

72 8

F

6

77 8

F

72/673/4

78 7

0.0 09±0.002

5

77 6

F

72/673/4 77

4

F

5

3 43/394/5 61

0.02± 0.0005

0.0 2±0.004

0.02±

0.0

07±0.001

04±0.001

0.0 08±0.001

09±0.001

u o

J

0.006 ±0.0005

l a

rn 0.

02±0.007

0.007

02±0.002

0.

02±0.007

0. 02±0.007

0.

02±0.002

0. 02±0.002

0. 02±0.002

8 0

08±0.009

5 1

7

p e 0.

06±0.008

05±0.008

5 2.4±5.8

6 3

8.8±12.8

2 01±30.4

1 0.4±0.6

M AM-3

06±29.4

0.8±0.4

AM-4

2

1

M

8

48±40.6

1.6±0.5

AM

2

1

C

5

7.5±2.3

2.6±0.7

AM

1

1

C

2

9.9±2.7

.48±0.1

AM

1

1

C

3

6.7±3.2

0

M

3

2.5±3.0

6

0. 06±0.008

8

6

0. 06±0.008

f o

1 .40±0.1

AM-4

9.2±5.7

5

0.

9

ro 3

M AM-4

4.8±10.5

0. 05±0.008

7 6.9±8.4

0.

r P 0.

02±0.007

±0.002

0.0

02±0.002

0.

0.008

08±0.008

0.

02±0.007

±0.0003

±0.0005

03±0.003

0.

0.007

0.

0.

02±0.007

±0.0002

0.0

03±0.003

0.

0.003

0.0

0.

02±0.007

0.0005

05±0.001

0.

1 80±26.4

0 .38±0.1

6. 69±2.0

55

Journal Pre-proof Table 3: Dose rate data, equivalent doses (De), overdispersion values, and OSL ages for lacustrine samples from Fern Gully Lagoon, NSI. The final OSL age of each sample has been calculated by dividing the De value by the total dose rate. a

Long-term water contents used for beta / gamma / cosmic-ray dose rate attenuation, expressed as % of dry mass of mineral fraction, with an assigned relative uncertainty

of ±10%. The final beta dose rates have been adjusted for moisture attenuation using the average water content from the midpoint of each OSL sample depth. The final gamma dose rates have been adjusted using the water content determined separately for the gamma dose rate bulk sediment samples, which were collected for each OSL sample depth, as well as for the overlying and underlying 10 cm depth. The final cosmic-ray dose rates have been adjusted using the average water content measured from

f o

the contiguous 1 cm3 bulk sediment samples collected throughout the overlying core sequence. b

o r p

Beta, gamma and internal dose rates have been calculated on dried and powdered sediment samples using ICP-MS and ICP-OES. The beta dose rates have been calculated

on bulk sediment samples collected from each OSL sample depth. The gamma dose rates have been determined separately on bulk sediment samples collected for each

e

OSL sample depth, as well as for the overlying and underlying 10 cm depth of each OSL sample position, following De Deckker et al. (2019). c

r P

Radionuclide concentrations have been converted to alpha, beta and gamma dose rates using the published conversion factors of Guérin et al. (2011), and allowing for

beta-dose attenuation (Mejdahl, 1979; Brennan, 2003) and long-term water content correction (Aitken, 1985). d

l a

An internal dose rate of 0.02 ± 0.007 Gy/kyr has been included in the final dose rate calculations of all samples, based on ICP-MS U and Th measurements made on

n r u

etched quartz grains from associated aeolian deposits at Welsby Lagoon (Lewis et al., in prep) and an alpha efficiency factor (a value) of 0.04 ± 0.01 (Rees-Jones, 1995; ReesJones and Tite, 1997).

o J

e

Cosmic-ray dose rates were calculated after Prescott and Hutton (1994), and assigned a relative uncertainty of ±10%.

f

Mean ± total uncertainty (68% confidence interval), calculated as the quadratic sum of the random and systematic uncertainties.

g

SG OSL = single-grain optically stimulated luminescence; MAM-3 = three-parameter minimum age model (Arnold et al., 2009), MAM-4 = four-parameter minimum

age model (Arnold et al., 2009); CAM = Central age model (Galbraith et al., 1999). MAM-3 and MAM-4 De estimates were calculated after adding, in quadrature, a relative error of 25% to each individual De measurement error to approximate the underlying dose overdispersion observed in an ‘ideal’ (well-bleached and unmixed) sedimentary sample from this core (FG15 OSL-3), which is consistent with global overdispersion datasets (Arnold and Roberts, 2009).

56

Journal Pre-proof h

Number of De measurements that passed the SAR quality assurance criteria and were used for De determination.

i

The relative spread in the De dataset beyond that associated with the measurement uncertainties for individual De values, calculated using the CAM.

j

Total uncertainty includes a systematic component of ±2% associated with laboratory beta-source calibration.

k

Samples excluded from the Bayesian model as they contain multiple discrete dose populations when fitted with the FMM (see section 3.5 for details).

f o

l a

o r p

r P

e

n r u

o J

57

Jo

ur

na

lP

re

-p

ro

of

Journal Pre-proof

Figure 6: Bayesian age/depth model for the Fern Gully Lagoon sequence obtained using a non-continuous deposition scenario. The prior age distributions for the dating samples (likelihoods) are shown in light blue. The modelled posterior 58

Journal Pre-proof distributions for the dating samples and unit boundaries are shown in dark blue and grey, respectively. Likelihood and posterior ages are shown on a calendar year timescale and expressed in years before 1950 AD. The 68.2% and 95.4%

Jo

ur

na

lP

re

-p

ro

of

ranges of the posterior probabilities are indicated by the light and dark shading

59

Journal Pre-proof

f o

l a

o r p

r P

e

n r u

Jo

60

Journal Pre-proof Figure 7: PCA biplot of µXRF elements and organic content from Fern Gully Lagoon. Crosses indicate the distribution mean and one standard deviation of the two principal axes of each distribution, while the outer polygon indicates the extent of the samples. Local inorganic source materials are indicated by grey points and grey arrows where these points lie outside the plot. Al.: aluminium; Ti.: titanium; K.: potassium; Si.: silicon; Zr.: zirconium.

f o

l a

o r p

r P

e

n r u

Jo

61

Jo

ur

na

lP

re

-p

ro

of

Journal Pre-proof

Figure 8: Sedimentation rate (3 cm moving average), sediment density (X-radiograph), inorganic content, magnetic susceptibility, calibrated µXRF aeolian input (silicon, potassium, zircon and titanium, PC1) and other calibrated µXRF inorganic sedimentation (calcium and iron, PC2). Records for comparison are Native Companion Lagoon

62

Journal Pre-proof (NCL) (Petherick et al., 2008), an Australian dust record from the South Pacific (Lamy et al., 2014), a global sea level record from the Red Sea (Rohling et al., 2009) and

Jo

ur

na

lP

re

-p

ro

of

December mean insolation for 30 degrees south (Berger and Loutre, 1991).

63

Journal Pre-proof

jgv

References:

Aitken, M.J., 1985. Thermoluminescence dating. Academic press.

Arnold, L.J., Roberts, R.G., 2009. Stochastic modelling of multi-grain equivalent dose (De) distributions: Implications for OSL dating of sediment mixtures. Quaternary Geochronology 4, 204-230.

of

Arnold, L.J., Roberts, R.G., Galbraith, R.F., DeLong, S.B., 2009. A revised burial dose estimation procedure for optical dating of young and modern-age sediments. Quaternary Geochronology 4, 306-

-p

ro

325.

lP

Quaternary Sciences Reviews 10, 297-317.

re

Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years.

BOM, 2005. Australian Köppen climate classifications based on a standard 30 year climatology

na

(1961-1990). Bureau of Meterology, Commonwealth of Australia.

ur

Bostock, H.C., Opdyke, B.N., Gagan, M.K., Kiss, A.E., Fifield, L.K., 2006. Glacial/interglacial

Jo

changes in the East Australian current. Climate Dynamics 26, 645-659.

Brennan, B.J., 2003. Beta doses to spherical grains. Radiation Measurements 37, 299-303.

De Deckker, P., Arnold, L.J., van der Kaars, S., Bayon, G., Stuut, J.-B.W., Perner, K., Lopes dos Santos, R., Uemura, R., Demuro, M., 2019. Marine Isotope Stage 4 in Australasia: A full glacial culminating 65,000 years ago – Global connections and implications for human dispersal. Quaternary Science Reviews 204, 187-207.

64

Journal Pre-proof Duller, G., 2015. The Analyst software package for luminescence data: overview and recent improvements. Ancient TL 33 p35-42.

Gagan, M.K., Ayliffe, L.K., Beck, J.W., Cole, J.E., Druffel, E.R.M., Dunbar, R.B., Schrag, D.P., 2000. New views of tropical paleoclimates from corals. Quaternary Science Reviews 19, 45-64.

Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M., 1999. Optical dating of single

of

and multiple grains of quartz from Jinmium Rock Shelter, northern Australia: Part 1, experimental design

ro

and statistical models. Archaeometry 41, 339-364.

-p

Guérin, G., Mercier, N., Adamiec, G., 2011. Dose-rate conversion factors: update. Ancient TL 29,

re

5-8.

lP

Jacobs, Z., Duller, G.A.T., Wintle, A.G., 2006. Interpretation of single grain De distributions and

na

calculation of De. Radiation Measurements 41, 264-277.

Lamy, F., Gersonde, R., Winckler, G., Esper, O., Jaeschke, A., Kuhn, G., Ullermann, J., Martinez-

ur

Garcia, A., Lambert, F., Kilian, R., 2014. Increased dust deposition in the Pacific Southern Ocean during

Jo

glacial periods. Science 343, 403-407.

Mejdahl, V., 1979. Thermoluminescence Dating: Beta-Dose Attenuation in Quartz Grains. Archaeometry 21, 61-72.

Petherick, L., McGowan, H., Moss, P., 2008. Climate variability during the Last Glacial Maximum in eastern Australia: evidence of two stadials? Journal of Quaternary Science Review 23, 787-802.

Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: Large depths and long-term time variations. Radiation Measurements 23, 497-500. 65

Journal Pre-proof Rees-Jones, J., 1995. Optical dating of young sediments using fine-grain quartz. Ancient TL 13, 914.

Rees-Jones, J., Tite, M.S., 1997. Optical dating results for British archaeological sediments. Archaeometry 36, 177-187.

Rohling, E.J., Grant, K., Bolshaw, M., Roberts, A.P., Siddall, M., Hemleben, C., Kucera, M., 2009.

of

Antarctic temperature and global sea level closely coupled over the past five glacial cycles. Nature

Jo

ur

na

lP

re

-p

ro

Geoscience 2, 500.

66

Journal Pre-proof

ur

na

lP

re

-p

ro

of

Records of Australian climate during Marine Isotope Stages 5 and 7 are rare. This high-resolution record indicates a wet MIS 7a–c before a relatively dry MIS 5. Inorganic flux during the Holocene was up to three times greater than during MIS 5. Findings are consistent with long-term drying observed in other regions. A review of µXRF calibration methods indicates a lack of standardisation.

Jo

    

67